JP6300884B2 - Image decoding apparatus, image decoding method, image encoding apparatus, image encoding method, and recording medium - Google Patents

Description

  The present invention relates to an image encoding apparatus that encodes an image and generates encoded data. The present invention also relates to an image decoding apparatus that decodes encoded data generated using such an image encoding apparatus.

  In order to efficiently transmit or record a moving image, a moving image encoding device is used. As a specific moving picture encoding method, for example, H.264 is used. H.264 / MPEG-4 AVC (Non-Patent Document 1) and a method employed in KTA software which is a codec for joint development in VCEG (Video Coding Expert Group).

  In such an encoding method, an image (picture) constituting a moving image is obtained by dividing a slice obtained by dividing an image, a macroblock obtained by dividing the slice, and a macroblock. Are managed by a hierarchical structure composed of sub-blocks, and are usually encoded for each sub-block.

  In such an encoding method, a predicted image is usually generated based on a local decoded image obtained by encoding / decoding an input image, and difference data between the predicted image and the input image is encoded. It becomes. As methods for generating a predicted image, methods called inter-frame prediction (inter prediction) and intra-frame prediction (intra prediction) are known.

  In inter prediction, a prediction image in a prediction target frame is generated by applying motion compensation using a motion vector to a reference image in a reference frame in which the entire frame is decoded. In inter prediction, it is also possible to generate a prediction image by referring to a plurality of reference images. In this case, the prediction image is obtained using a value obtained by multiplying the pixel value of each reference image by a weighting coefficient. Generated.

  On the other hand, in intra prediction, based on locally decoded images in the same frame, predicted images in the frame are sequentially generated. Specifically, in intra prediction, usually, prediction directions included in a predetermined prediction direction (prediction mode) group for each prediction unit (for example, subblock) constituting a unit region (for example, macroblock). Is selected, and the pixel value of the reference pixel in the locally decoded image is extrapolated to the selected prediction direction, thereby generating a prediction pixel value on the prediction target region.

  As described above, the predicted image is generally generated based on a prediction parameter such as a motion vector, a weighting coefficient, or a prediction mode.

ITU-T Recommendation H. H.264 (11/07) (released in November 2007)

However, in order to appropriately generate a predicted image in the video decoding device, it is necessary to encode the prediction parameter used in the video encoding device and transmit it to the video decoding device. There is a problem that the code amount of data increases.

  For example, in the above-described conventional intra prediction, it is necessary to encode prediction mode information that is information indicating which prediction mode has been selected for each prediction target region together with the difference data. There is a problem that the amount of encoded data increases due to the prediction mode information.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide an image encoding device capable of reducing the amount of code for specifying a prediction parameter without sacrificing encoding efficiency, Another object of the present invention is to realize an image decoding apparatus capable of decoding encoded data generated by such an image encoding apparatus.

  In order to solve the above problem, an image decoding apparatus according to an aspect of the present invention encodes a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit. In an image decoding apparatus for decoding the obtained encoded data, a prediction parameter decoding unit that decodes a prediction parameter applied to each prediction unit, a reduced set construction unit that generates a reduced set that is a set of prediction parameters, and a basic set Decoding the first prediction parameter that identifies the prediction parameter included in the length, or determining the second prediction parameter that identifies the prediction parameter included in the reduced set according to the number of prediction parameters included in the reduced set 2nd prediction parameter decoding means for decoding from the bit string of the above, the reduced set is the target prediction unit And the number of prediction parameters included in the reduced set is less than the number of prediction parameters included in a basic set of predetermined prediction parameters, and includes at least one prediction parameter included in the neighboring prediction unit region The prediction parameter decoded by the parameter decoding unit is the first prediction parameter decoded by the second prediction parameter decoding unit according to the value of the flag included in the encoded data, or the second prediction parameter decoding unit. The second prediction parameter decoded according to the value of the flag included in the encoded data is included, and the prediction parameter is a motion vector used for motion compensation prediction and is included in the neighborhood prediction unit region for the target prediction unit. Between at least two prediction parameters If there are multiple prediction parameters, one of the overlapping prediction parameters is included in the reduced set, the prediction parameter decoding means decodes the prediction parameter included in the reduced set, and the prediction parameter The decoding means is characterized in that when the number related to the prediction parameter included in the reduced set is 1, decoding of the prediction parameter is performed by omitting decoding of information regarding the prediction parameter.

  In order to solve the above problem, an image decoding method according to an aspect of the present invention encodes a difference between an input image and a predicted image together with a prediction parameter used to generate a predicted image for each prediction unit. An image decoding method for decoding the obtained encoded data, a prediction parameter decoding step for decoding a prediction parameter applied to each prediction unit, a reduced set construction step for generating a reduced set that is a set of prediction parameters, A first prediction parameter specifying a prediction parameter included in the basic set is decoded, or a second prediction parameter specifying a prediction parameter included in the reduced set is determined based on the number of prediction parameters included in the reduced set. A second prediction parameter decoding step for decoding from a bit string of length, the reduced set comprising: Including at least one prediction parameter included in the neighborhood prediction unit region for the elephant prediction unit, and the number of prediction parameters included in the reduced set is less than the number of prediction parameters included in a basic set of prediction parameters determined in advance. The prediction parameter decoded in the prediction parameter decoding step is the first prediction parameter decoded according to the value of the flag included in the encoded data in the second prediction parameter decoding step, or the second prediction parameter The second prediction parameter decoded according to the value of the flag included in the encoded data in the decoding step is included, and the prediction parameter is a motion vector used for motion compensation prediction, and is a neighborhood prediction unit for the target prediction unit At least two or more prediction parameters included in the region If there is a prediction parameter that overlaps between parameters, one of the overlapping prediction parameters is included in the reduced set, and in the prediction parameter decoding step, the prediction parameter included in the reduced set is decoded, In the prediction parameter decoding step, when the number related to the prediction parameters included in the reduced set is 1, decoding of the information related to the prediction parameters is omitted and the prediction parameters are decoded.

  In order to solve the above problem, an image encoding device according to an aspect of the present invention encodes a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit. In the image encoding device that generates encoded data according to the above, a prediction parameter encoding unit that encodes a prediction parameter applied to each prediction unit, a reduced set construction unit that generates a reduced set that is a set of prediction parameters, and a basic A first prediction parameter specifying a prediction parameter included in the set is encoded, or a second prediction parameter specifying a prediction parameter included in the reduced set is determined by the number of prediction parameters included in the reduced set. Second prediction parameter encoding means for encoding from a bit string of length, and the reduced set includes a target prediction. Including at least one or more prediction parameters included in the neighborhood prediction unit region for the unit, and the number of prediction parameters included in the reduced set is less than the number of prediction parameters included in a basic set of predetermined prediction parameters, The prediction parameter encoded by the prediction parameter encoding means is the first prediction parameter encoded according to the value of the flag included in the encoded data encoded by the second prediction parameter encoding means, or The second prediction parameter encoded according to the value of the flag included in the encoded data encoded by the second prediction parameter encoding means is included, and the prediction parameter is used for motion compensation prediction. A motion vector, at least 2 included in the neighborhood prediction unit region for the target prediction unit If there is a prediction parameter that overlaps between the prediction parameters in the upper prediction parameter, one of the overlapping prediction parameters is included in the reduced set, and the prediction parameter encoding means includes the reduced parameter in the reduced set. When the number of prediction parameters included in the reduced set is 1, the prediction parameter encoding means encodes the prediction parameters by omitting encoding of information regarding the prediction parameters. It is characterized by becoming.

  In order to solve the above problem, an image encoding method according to an aspect of the present invention encodes a difference between an input image and a predicted image together with a prediction parameter used to generate a predicted image for each prediction unit. In the image encoding method for generating encoded data by the above, a reduced set construction step for generating a reduced set that is a set of prediction parameters applied to each prediction unit, and a first prediction parameter for specifying a prediction parameter included in the basic set Or a second prediction parameter code that encodes a second prediction parameter that identifies a prediction parameter included in the reduced set from a bit string having a length determined by the number of prediction parameters included in the reduced set And a prediction parameter encoding step for encoding a prediction parameter, and Includes at least one prediction parameter included in the neighborhood prediction unit area for the target prediction unit, and the number of prediction parameters included in the reduced set is the number of prediction parameters included in a basic set of predetermined prediction parameters. The first prediction parameter encoded in the prediction parameter encoding step is encoded according to the value of the flag included in the encoded data encoded in the second prediction parameter encoding step. A prediction parameter, or a second prediction parameter that is encoded according to a flag value included in the encoded data encoded by the second prediction parameter encoding step, and the prediction parameter is A motion vector used for motion compensated prediction, and a neighborhood prediction unit for the target prediction unit When at least two or more prediction parameters included in a region include prediction parameters that overlap between prediction parameters, one of the overlapping prediction parameters is included in the reduced set, and the prediction parameter encoding step Then, the prediction parameter included in the reduced set is encoded. In the prediction parameter encoding step, when the number related to the prediction parameter included in the reduced set is 1, the encoding of information regarding the prediction parameter is omitted. Thus, the prediction parameter is encoded.

  In order to solve the above problems, a recording medium according to an aspect of the present invention obtains a difference between an input image and a predicted image by encoding each prediction unit together with a prediction parameter used to generate a predicted image. A computer-readable recording medium recording a control program for causing a computer to execute an image decoding method for decoding the encoded data, and a prediction parameter decoding step for decoding a prediction parameter to be applied to each prediction unit; A reduced set construction step for generating a reduced set that is a set of prediction parameters, and a first prediction parameter that specifies a prediction parameter included in the basic set, or a second that specifies a prediction parameter included in the reduced set The prediction parameter is determined by the number of prediction parameters included in the reduced set. A second prediction parameter decoding step for decoding from a bit string of length, wherein the reduced set includes at least one or more prediction parameters included in a neighboring prediction unit region for a target prediction unit, and includes a prediction included in the reduced set The number of parameters is less than the number of prediction parameters included in a basic set of predetermined prediction parameters, and the prediction parameters decoded in the prediction parameter decoding step are included in the encoded data in the second prediction parameter decoding step. The first prediction parameter decoded according to the value of the flag to be decoded, or the second prediction parameter decoded according to the value of the flag included in the encoded data in the second prediction parameter decoding step. The prediction parameter is a motion vector used for motion compensation prediction. If at least two or more prediction parameters included in the neighboring prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. In addition, in the prediction parameter decoding step, the prediction parameter included in the reduced set is decoded. In the prediction parameter decoding step, if the number related to the prediction parameter included in the reduced set is 1, information on the prediction parameter This is a computer-readable recording medium on which a control program for causing a computer to execute the decoding of the prediction parameter without decoding is recorded.

  In order to solve the above problems, a recording medium according to an aspect of the present invention encodes a difference between an input image and a prediction image by encoding the prediction unit for each prediction unit together with a prediction parameter used to generate a prediction image. A computer-readable recording medium storing a control program for causing a computer to execute an image encoding method for generating encoded data, and generating a reduced set that is a set of prediction parameters applied to each prediction unit A first prediction parameter for specifying a prediction parameter included in the construction step and the basic set, or a second prediction parameter for specifying a prediction parameter included in the reduced set is used as a prediction parameter included in the reduced set. A second prediction parameter encoding step encoded from a bit string having a length determined by the number of And a prediction parameter encoding step for encoding a prediction parameter, and the reduced set includes at least one or more prediction parameters included in a neighboring prediction unit region for the target prediction unit and is included in the reduced set. The number of prediction parameters is less than the number of prediction parameters included in a basic set of predetermined prediction parameters, and the prediction parameter encoded in the prediction parameter encoding step is encoded by the second prediction parameter encoding step. The first prediction parameter encoded according to the flag value included in the encoded data or the flag value included in the encoded data encoded by the second prediction parameter encoding step. A second prediction parameter encoded according to A parameter is a motion vector used for motion compensation prediction, and when there is a prediction parameter that overlaps between the prediction parameters in at least two or more prediction parameters included in the neighborhood prediction unit region with respect to the target prediction unit, the overlapping prediction is performed. One of the parameters is included in the reduced set, the prediction parameter encoding step encodes a prediction parameter included in the reduced set, and the prediction parameter encoding step relates to a prediction parameter included in the reduced set. When the number is 1, it is a computer-readable recording medium on which a control program for causing a computer to execute encoding of the prediction parameter by omitting encoding of information regarding the prediction parameter is recorded.

  According to the image coding apparatus configured as described above, it is possible to reduce the code amount for designating the prediction parameter without sacrificing the coding efficiency.

It is a block diagram which shows the structure of the moving image decoding apparatus which concerns on embodiment. It is a block diagram which shows the structure of the MB decoding part with which the moving image decoding apparatus which concerns on embodiment is provided. It is a block diagram which shows the structure of the prediction parameter decoding part with which the moving image decoding apparatus which concerns on embodiment is provided. It is a figure for demonstrating operation | movement of the group determination part with which a prediction parameter decoding part is provided. (A)-(b) shows the case where each of 16 subblocks contained in a macroblock is classified into either the 1st group or the 2nd group based on classification method A (C) to (d) show the case where each sub-block is classified based on the classification method B, and (e) to (f) show that each sub-block is based on the classification method C. Shows the case of classification. H. It is a figure which shows the intra prediction mode used for the intra prediction in H.264 / MPEG-4 AVC standard with the index attached | subjected to each prediction mode. It is a figure for demonstrating operation | movement of the reduced set derivation | leading-out part with which a prediction parameter decoding part is provided. (A) is a flowchart which shows the 1st example of the production | generation operation | movement of the reduction | restoration set in a reduction | decrease set derivation part, (b) is a flowchart which shows the 2nd example of the production | generation operation | movement of the reduction | decrease set in a reduction | decrease set derivation part. Yes, (c) is a flowchart showing a third example of a reduced set generation operation in the reduced set derivation unit. It is a flowchart which shows an example of the flow of the decoding process in the 2nd prediction parameter decoding part with which a prediction parameter decoding part is provided. It is a figure for demonstrating the other structural example of a prediction parameter decoding part. (A) is a flowchart which shows the production | generation operation | movement of the reduction | restoration set by a reduction | restoration set derivation | leading-out part, (b) has shown an example of the neighborhood subblock area | region. It is for demonstrating the production | generation process of the prediction image by the prediction image generation part with which MB decoding part is provided, Comprising: Each pixel of the prediction object subblock which is 4x4 pixels, The surrounding pixel of the said prediction object subblock, FIG. It is a block diagram which shows the structure of the moving image encoder which concerns on embodiment. It is a block diagram which shows the structure of the MB encoding part with which the moving image encoder which concerns on embodiment is provided. It is a block diagram which shows the structure of the prediction parameter determination part with which an MB encoding part is provided. It is a figure for demonstrating operation | movement of the prediction parameter determination part with which an MB encoding part is provided. (A) has shown the example of the prediction mode which the 1st prediction parameter determination part selected with respect to each subblock which belongs to a 1st group among each subblock which comprises macroblock MB, (b) Shows an example of a reduced set generated by the reduced set derivation unit when each prediction mode shown in (a) is supplied as a prediction parameter, and (c) shows a second prediction parameter determination unit, The example of the prediction mode selected from the reduction | decrease set shown to (b) with respect to each subblock which belongs to a 2nd group is shown. It is a block diagram which shows the structure of the prediction parameter encoding part with which an MB encoding part is provided. It is a figure which shows the bit stream structure for every macroblock of the encoding data produced | generated by the moving image encoder which concerns on embodiment, and is referred in the moving image decoder which concerns on embodiment. It is a figure which shows the other example of a basic parameter set. (A) shows an example of a parameter set that emphasizes the horizontal direction, and (b) shows an example of a parameter set that emphasizes the vertical direction.

(Video decoding device)
A configuration of a moving image decoding apparatus (image decoding apparatus) 1 according to the embodiment will be described with reference to FIGS. The moving picture decoding apparatus 1 includes H.264 as a part thereof. This is a moving picture coding apparatus using a technique adopted in the H.264 / MPEG-4 AVC standard.

  Generally speaking, the moving image decoding apparatus 1 is an apparatus that generates and outputs a decoded image # 2 by decoding input encoded data # 1.

  In addition, the video decoding device 1 divides a unit region on the image indicated by the encoded data # 1 into a plurality of prediction target regions (prediction units), and uses the prediction image generated for each prediction target region, Decoded image # 2 is generated.

  In the following, the unit area is referred to as H.264. The present invention will be described with reference to an example in which a macroblock in the H.264 / MPEG-4 AVC standard is used and the prediction target area is a subblock in the macroblock, but the present invention is not limited to this. For example, the unit area may be an area larger than a macroblock, or may be an area overlapping with a plurality of macroblocks.

  FIG. 1 is a block diagram showing a configuration of the moving picture decoding apparatus 1. As shown in FIG. 1, the moving picture decoding apparatus 1 includes a variable length demultiplexing unit 11, a header information decoding unit 12, an MB setting unit 13, an MB decoding unit 14, and a frame memory 15.

  The encoded data # 1 input to the video decoding device 1 is input to the variable length demultiplexing unit 11. The variable length demultiplexing unit 11 demultiplexes the input encoded data # 1, thereby converting the encoded data # 1 into header encoded data # 11a that is encoded data related to header information, and a macroblock Separated into MB encoded data # 11b, which is encoded data relating to (unit area), header encoded data # 11a is output to header information decoding unit 12, and MB encoded data # 11b is output to MB setting unit 13, respectively. .

  The header information decoding unit 12 decodes the header information # 12 from the encoded header data # 11a. Here, the header information # 12 is information including the size of the input image.

  The MB setting unit 13 separates the MB encoded data # 11b into encoded data # 13 corresponding to each macroblock based on the input header information # 12, and sequentially outputs the encoded data # 11b to the MB decoding unit 14. .

  The MB decoding unit 14 generates and outputs a decoded image # 2 corresponding to each macroblock by sequentially decoding the encoded data # 13 corresponding to each input macroblock. The decoded image # 2 is also output to the frame memory 15. The configuration of the MB decoding unit 14 will be described later and will not be described here.

The frame memory 15 records the decoded image # 2. In the frame memory 15, at the time of decoding a specific macroblock, decoded images corresponding to all the macroblocks that precede the macroblock in the raster scan order are recorded.

  When all the macroblocks in the image have been decoded by the MB decoding unit 14 in units of macroblocks, the decoded image # 2 corresponding to the encoded data input to the video decoding device 1 The generation process is completed.

(MB decoding unit 14)
Hereinafter, the MB decoding unit 14 will be described more specifically with reference to different drawings.

  FIG. 2 is a block diagram showing a configuration of the MB decoding unit 14. As shown in FIG. 2, the MB decoding unit 14 includes a sub-block dividing unit 141, a prediction residual decoding unit 142, a sub-block decoded image generation unit 143, a prediction parameter decoding unit 144, a prediction image generation unit 145, and MB decoding. An image generation unit 146 is provided.

  The sub-block dividing unit 141 is activated when encoded data # 13 in units of macro blocks is input, and each sub block (each prediction target region) constituting the macro block (unit region) within the corresponding macro block. Sub-block position information # 141a indicating the position and sub-block encoded data # 141b which is encoded data related to the sub-block indicated by the sub-block position information # 141a are sequentially output in a predetermined order. Note that the method used in the video encoding apparatus that generates the encoded data # 1 can be applied to the method of dividing the macroblock into sub-blocks.

  The sub-block dividing unit 141 outputs sub-block position information # 141a and sub-block encoded data # 141b related to sub-blocks belonging to the first group described later, and then outputs sub-blocks related to sub-blocks belonging to the second group described later. It is preferable that the position information # 141a and the sub-block encoded data # 141b are output. For example, it is preferable that the sub blocks belonging to the first group are scanned in the raster scan order, and then the sub blocks belonging to the second group are scanned in the raster scan order. Further, the sub-block dividing unit 141 outputs the sub-block position information # 141a and the sub-block encoded data # 141b in the same order as that used in the video encoding device that generates the encoded data # 1. You may make it do.

  The prediction residual decoding unit 142 generates transform coefficients for the subblock indicating the input subblock position information # 141a by applying variable length code decoding to the input subblock encoded data # 141b. Further, the prediction residual decoding unit 142 applies the inverse transform of DCT (Discrete Cosine Transform) having the same size as the size of the sub-block indicated by the sub-block position information # 141a to the generated transform coefficient. A difference # 142 is generated and output.

  Prediction parameter decoding section 144 decodes and outputs prediction parameter # 144 for each subblock based on subblock position information # 141a and subblock encoded data # 141b.

  Here, the prediction parameter refers to a parameter used for generating a predicted image. Examples of prediction parameters include a prediction mode in intra prediction, a motion vector in motion compensated prediction, a weight coefficient in luminance compensated prediction, and the like.

  The prediction parameter # 144 includes a prediction parameter # 43 output from a first prediction parameter decoding unit 43 described later and a prediction parameter # 45 output from a second prediction parameter decoding unit 45 described later. Since the specific configuration and operation of the prediction parameter decoding unit 144 will be described later, description thereof is omitted here.

  The predicted image generation unit 145 generates a predicted image # 145 corresponding to the prediction target sub-block based on the prediction parameter # 144, the decoded image # 2, and the decoded image # 15 recorded in the frame memory 15, Output. Since a specific method for generating the predicted image # 145 in the predicted image generation unit 145 will be described later, description thereof is omitted here.

  The sub-block decoded image generation unit 143 adds the prediction image # 145 to the decoding residual # 142 to generate and output a sub-block decoded image # 143 that is a decoded image in units of sub-blocks.

  The MB decoded image generation unit 146 accumulates the subblock decoded image # 143 in units of subblocks for each macroblock, and integrates all the subblock decoded images # 143 constituting the macroblock, thereby decoding the decoded image in units of macroblocks. # 2 is generated and output. The generated decoded image # 2 is also supplied to the predicted image generation unit 145.

(Prediction parameter decoding unit 144)
Next, the configuration of the prediction parameter decoding unit 144 will be described with reference to FIG.

  FIG. 3 is a block diagram illustrating a configuration of the prediction parameter decoding unit 144. As illustrated in FIG. 3, the prediction parameter decoding unit 144 includes a group determination unit 41, a switch unit 42, a first prediction parameter decoding unit 43, a reduced set derivation unit 44, and a second prediction parameter decoding unit 45. .

(Group determination unit 41)
The group determination unit 41 determines which group of a plurality of predetermined groups the subblock indicated by the subblock position information # 141a belongs to, and sets the group information # 41 indicating the determination result to the switch unit 42. Output for.

  Here, the predetermined plurality of groups refers to, for example, a plurality of groups into which each sub-block is classified in a moving image encoding apparatus that generates encoded data # 1. That is, in the moving picture encoding apparatus that generates the encoded data # 1, each of the subblocks SB1 to SBNs (Ns is the total number of subblocks belonging to the macroblock MB) belonging to a certain macroblock MB is determined as a predetermined classification method. Based on the group GP1 to GPM (M is the total number of groups into which the subblocks belonging to the macroblock MB are classified), the subblock SBn is classified into the group GPm. In this case, the group determination unit determines that the sub block SBn indicated by the sub block position information # 141a belongs to the group GPm based on the predetermined classification method, for example.

  Hereinafter, the case where each sub-block is classified into two groups will be described as an example with reference to (a) to (f) of FIG.

  4A to 4B, each of the 16 sub-blocks included in the macroblock MB is classified into either the first group or the second group based on the classification method A. 4 (c) to (d) are diagrams illustrating a case in which each sub-block is classified based on the classification method B, and (e) to (f) in FIG. FIG. 10 is a diagram illustrating a case where each sub-block is classified based on a classification method C.

As shown in FIGS. 4A to 4B, each subblock included in the macroblock MB is first checked so that the arrangement of the subblocks included in each group is a checkered flag.
Or the second group, and as shown in FIGS. 4C to 4D, the sub-blocks included in each group are classified so as to be adjacent only in the horizontal direction. Alternatively, as shown in FIGS. 4E to 4F, the sub-blocks included in each group may be classified so as to be adjacent only in the vertical direction.

  In general, the optimal classification method varies depending on the spatial correlation of the prediction parameters in the macroblock. The above classification method A is an effective classification method regardless of whether there is any spatial correlation in the vertical direction or the horizontal direction. On the other hand, when there is an oblique edge in the macroblock, the above classification method B and classification method C are effective.

  Regardless of which classification method is used, as is clear from (a) to (f) of FIG. 4, each sub-block is divided into the first group or the first group according to the position in the macroblock MB. It is classified into one of the second groups.

  The group determination unit 41 refers to the sub-block position information # 141a, and the sub-block indicated by the sub-block position information # 141a is determined based on the classification method used in the video encoding device that generates the encoded data # 1. Which group of the first group and the second group belongs is determined.

  For example, in the moving picture encoding apparatus that generates the encoded data # 1, the sub-block SB1 belonging to the macroblock MB is second based on the classification method A, as shown in FIGS. When the sub-block SB2 is classified into the first group, the group determination unit 41 refers to the sub-block position information # 141a based on the classification method A, so that the sub-block SB2 Are determined to belong to the first group, and the sub-block SB1 is determined to belong to the second group. The same applies to other sub-blocks included in the macro block MB.

  Also, in the video encoding device that generates the encoded data # 1, when a different classification method is used for each macroblock, the encoded data # 1 indicates which classification method is used for each macroblock. It is preferable to include a flag to indicate. By referring to such a flag, the group determination unit 41 can make a determination based on the classification method used in the video encoding device even when the classification method is different for each macroblock. it can.

  In the above description, the number of sub-blocks included in the macroblock is 16, but the present invention is not limited to this (the same applies hereinafter). The subblock classification method in the macroblock MB is not limited to the above-described example, and other classification methods may be used. For example, the classification method may be such that the number of subblocks belonging to the first group and the number of subblocks belonging to the second group are different from each other (the same applies hereinafter).

(Switch section 42)
Based on the group information # 41, the switch unit 42 converts the subblock encoded data # 141b, which is encoded data related to the subblock indicated by the subblock position information # 141a, to the first prediction parameter decoding unit 43 or the second The prediction parameter decoding unit 45 is transmitted to one of the parameter decoding units.

Specifically, when the group determination unit 41 determines that the subblock indicated by the subblock position information # 141a belongs to the first group, the switch unit 42 performs the subblock encoded data. # 141b is transmitted to the first prediction parameter decoding unit 43, and when the group determination unit 41 determines that the subblock indicated by the subblock position information # 141a belongs to the second group, The sub-block encoded data # 141b is transmitted to the second prediction parameter decoding unit 45.

(First prediction parameter decoding unit 43)
The first prediction parameter decoding unit 43 decodes the sub-block encoded data # 141b, thereby generating the encoded data # 1 in the moving picture encoding apparatus, and the sub-block (indicated by the sub-block position information # 141a ( The prediction parameter # 43 used for prediction of the prediction target sub-block is decoded and output.

  More specifically, the first prediction parameter decoding unit 43 first determines a prediction parameter used for prediction of the upper or left subblock of the prediction target subblock, and the decoded prediction parameter. Set to an estimate for the block.

  Subsequently, the first prediction parameter decoding unit 43 decodes the flag included in the sub-block encoded data # 141b.

  When the flag indicates that the estimated value is used, the estimated value is set as a prediction parameter for the prediction target sub-block, and the flag indicates that the estimated value is not used. Sets the prediction parameter decoded from the part other than the flag as the prediction parameter for the prediction target sub-block.

  If the upper or left sub-block of the prediction target sub-block is not decoded, the decoded sub-block above or to the left of the prediction target sub-block and the sub-block closest to the prediction target sub-block The prediction parameter used for prediction may be referred to as the estimated value.

  The decoded prediction parameter # 43 is also supplied to the reduced set derivation unit 44.

  Through the above operation, the reduced set derivation unit 44 is supplied with the prediction parameter # 43 decoded from each sub-block belonging to the first group.

(Reduced set derivation unit 44)
The reduced set deriving unit 44 accumulates the prediction parameter # 43 and generates a reduced prediction parameter set RS (hereinafter referred to as “reduced set RS”). Here, the reduced set RS is a set including the prediction parameter # 43 decoded from each sub-block belonging to the first group. Further, the reduced set RS may include a prediction parameter other than the prediction parameter # 43.

  When the same prediction parameter is decoded from a plurality of sub-blocks belonging to the first group, the reduced set derivation unit 44 includes only one of the same prediction parameter, A reduced set RS may be generated. In other words, the reduced set deriving unit 44 may generate the reduced set RS so that the prediction parameters do not overlap. For example, when the prediction parameter PP1 is decoded from each of the subblocks SB1 to SB8 among the subblocks SB1 to SB16 belonging to the first group, and the prediction parameter PP2 is decoded from each of the subblocks SB9 to SB16, The reduced set derivation unit 44 generates a reduced set RS that includes one prediction parameter PP1 and one prediction parameter PP2.

  In the following, the prediction parameter is H.264. The generation operation of the reduced set RS by the reduced set derivation unit 44 will be described with reference to FIGS. 5 and 6A to 6C, taking the case of the intra prediction mode in the H.264 / MPEG-4 AVC standard as an example. To do.

  FIG. 2 is a diagram illustrating an intra prediction mode (hereinafter referred to as “prediction mode”) used for intra prediction in the H.264 / MPEG-4 AVC standard, and indexes assigned to the respective prediction modes. Each intra prediction mode represents a prediction direction used for intra prediction, and as shown in FIG. In the H.264 / MPEG-4 AVC standard, eight-direction prediction modes (corresponding to indexes 0, 1, 3 to 8) and DC prediction modes (corresponding to index 2) are used. Hereinafter, the prediction mode specified by the index I is represented as the prediction mode I. In addition, a parameter set including the prediction mode 0 to the prediction mode 8 is referred to as a basic parameter set.

(Generation example 1 of the reduced set RS)
FIG. 6A is a flowchart illustrating a first example of the generation operation of the reduced set RS in the reduced set derivation unit 44.

  As shown in FIG. 6A, first, the reduced set derivation unit 44 initializes the reduced set RS by setting the reduced set RS to be empty (step S101).

  Subsequently, the reduced set deriving unit 44 adds the prediction parameter # 43 decoded from each sub-block belonging to the first group to the reduced set RS (step S102). For example, when prediction mode 1, prediction mode 6, and prediction mode 8 are decoded from each sub-block belonging to the first group, the reduced set derivation unit 44 includes prediction mode 1, prediction mode 6, and Prediction mode 8 is added to the reduced set RS.

  By performing the above operation, in the first example, the reduced set derivation unit 44 generates a reduced set RS configured by the prediction parameter # 43 decoded from each sub-block belonging to the first group. Can do.

  In general, there is a correlation between optimal prediction parameters for each sub-block constituting a macroblock. Therefore, the prediction parameter selected for each sub-block belonging to the first group is highly likely to be an optimal prediction parameter for each sub-block belonging to the second group. Further, the number of prediction modes included in the reduced set RS is smaller than the number of prediction parameters included in the basic parameter set.

  Therefore, the moving picture encoding apparatus that generates the encoded data # 1 takes the configuration corresponding to the configuration of the present example, so that the encoded data # 1 having a small code amount can be obtained without sacrificing the encoding efficiency. Can be generated. Also, the moving picture decoding apparatus 1 can decode the encoded data # 1 generated in this way with a small code amount by adopting the configuration of this example.

(Generation example 2 of reduced set RS)
FIG. 6B is a flowchart illustrating a second example of the generation operation of the reduced set RS in the reduced set derivation unit 44.

  As shown in FIG. 6B, first, the reduced set derivation unit 44 initializes the reduced set RS by setting the reduced set RS to be empty (step S201).

  Subsequently, the reduced set deriving unit 44 adds the additional parameter set AS to the reduced set RS (step S202). Here, it is preferable that the additional parameter set AS includes a prediction parameter that tends to be frequently used. In general, H.W. In the prediction mode of the H.264 / MPEG-4 AVC standard, a prediction mode specified by a smaller index tends to be used more frequently in intra prediction. Preferably a mode is included. For example, the additional parameter set AS is preferably configured to include a prediction mode 0 (vertical prediction mode), a prediction mode 1 (horizontal prediction mode), and a prediction mode 2 (DC prediction mode).

  Further, the additional parameter set AS may be configured to include at least one prediction mode among the prediction mode 0, the prediction mode 1, and the prediction mode 2.

  Subsequently, the reduced set deriving unit 44 adds the prediction parameter # 43 decoded from each sub-block belonging to the first group to the reduced set RS (step S203). However, it is preferable that the reduced set derivation unit 44 does not add the prediction parameters already included in the reduced set RS of the prediction parameters # 43 to the reduced set RS in order to avoid duplication. For example, when the prediction parameter # 43 is the prediction mode 1 and the prediction mode 4, and the prediction mode 1 and the prediction mode 2 are already included in the reduced set RS, the reduced set derivation unit 44 Need only add prediction mode 4 to the reduced set RS.

  By performing the above operation, in the second example, the reduced set deriving unit 44 has the prediction parameter # 43 decoded from each sub-block belonging to the first group, the prediction mode included in the additional parameter set, and A reduced set RS can be generated.

  By constructing a reduced set RS as described above, a reduced set RS composed of prediction parameters # 43 decoded from each sub-block belonging to the first group and a prediction mode that tends to be frequently used is obtained. Can be generated.

  Therefore, the moving picture coding apparatus that generates the coded data # 1 and includes the reduced set derivation unit that operates as in this example is a code that has a smaller code amount of the prediction residual. Data # 1 can be generated. Also, the moving picture decoding apparatus 1 including the reduced set deriving unit 44 that operates as in the present example can decode the encoded data # 1 having a smaller code amount of such a prediction residual.

(Generation example 3 of the reduced set RS)
FIG. 6C is a flowchart illustrating a third example of the generation operation of the reduced set RS in the reduced set derivation unit 44.

  As shown in FIG. 6C, first, the reduced set derivation unit 44 initializes the reduced set RS by setting the reduced set RS to be empty (step S301).

  Subsequently, the reduced set deriving unit 44 adds the prediction parameter # 43 decoded from each sub-block belonging to the first group to the reduced set RS (step S302).

  Subsequently, the reduced set deriving unit 44 determines whether or not log2 (Np−1) is an integer (step S303). Here, Np is the number of prediction parameters included in the reduced set RS.

  When log2 (Np−1) is an integer (Yes in step S303), the reduced set deriving unit 44 outputs the reduced set RS.

  When log2 (Np−1) is not an integer (No in step S303), the reduced set deriving unit 44 adds a predetermined prediction parameter to the reduced set (step S304), and performs the process of step S303 again. Here, as the predetermined prediction parameter, for example, among the prediction modes 0 to 8 included in the basic parameter set, a prediction mode that is not included in the reduced set RS and has the smallest index is selected. Just do it.

  As described above, the prediction mode specified by a smaller index tends to be used more frequently in intra prediction. Therefore, the reduced set deriving unit 44 adds a prediction mode that tends to be used more frequently in intra prediction to the reduced set in this step.

  By performing the above operation, in the third example, the reduced set deriving unit 44 is a reduced set including the prediction parameter # 43 decoded from each sub-block belonging to the first group, and 2n + 1 ( It is possible to generate a reduced set RS including n prediction parameters.

  In general, when the prediction parameter is variable-length encoded together with a flag indicating whether or not it is the same as the estimated value, the number of prediction parameters is 2n + 1 (n is an integer), compared to the case where it is not. Therefore, there is a tendency that the compression efficiency when the prediction parameter is variable-length encoded increases.

  Therefore, the reduced set deriving unit 44 can generate a reduced set RS with high compression efficiency when performing variable length coding by performing the above-described operation. Therefore, a moving image encoding apparatus that generates encoded data # 1 and that includes a reduced set derivation unit that operates as in this example generates encoded data # 1 with high compression efficiency. can do. In addition, the moving picture decoding apparatus 1 including the reduced set deriving unit 44 that operates as in the present example can decode such encoded data # 1 with high compression efficiency.

  If the number of prediction parameters # 43 is not 2n + 1 (n is an integer), the reduced set deriving unit 44 can generate the reduced set RS so as to include the predetermined prediction parameter. A reduced set RS including a prediction mode that tends to be frequently used can be generated.

(Generation example 4 of reduced set RS)
In any of the three generation examples of the reduced set RS shown in (a) to (c) of FIG. 6, the reduced set derivation unit 44 predicts the prediction parameter # decoded from each sub-block belonging to the first group. Of 43, all kinds of prediction parameters except for overlapping ones are added to the reduced set RS. However, the prediction parameters decoded from the sub-blocks belonging to the first group may be configured to add only some types instead of all types.

  Specifically, a configuration may be adopted in which only the prediction parameters whose appearance ratio is higher than a predetermined value in the set of prediction parameters decoded from the sub-blocks belonging to the first group are added to the reduced set RS. Here, the appearance ratio of the prediction parameter is, for example, dividing the number of subblocks to which the prediction parameter is assigned among the subblocks belonging to the first group by the number of all subblocks belonging to the first group. Can be defined by For example, if the number of all subblocks belonging to the first group is Nf, and the number of subblocks to which the prediction parameter Pa is decoded and allocated among the subblocks belonging to the first group is Npa, the prediction parameter The appearance ratio of P can be defined by Npa / Nf. The appearance ratio can also be expressed as a percentage.

  Further, this generation example will be described in more detail as follows when eight sub-blocks (sub-blocks SB1 to SB8) belong to the first group.

  For example, prediction mode 0 is decoded for sub-blocks SB1, SB2, SB3, and SB4, prediction mode 1 is decoded for sub-blocks SB5 and SB6, and prediction mode 2 is decoded for sub-block SB7. When the prediction mode 3 is decoded for the block SB8 and the predetermined value is set to 40%, only the prediction mode 0 having the appearance ratio of 50% is added to the reduced set RS. The On the other hand, when the predetermined value is set to 20%, the prediction mode 0 in which the appearance ratio is 50% and the prediction mode 1 in which the appearance ratio is 25% are added to the reduced set RS. .

  In general, when the number of sub-blocks included in a macroblock is large, it may be difficult to effectively reduce the code amount by including a large number of prediction parameters in the reduced set RS. .

  According to this generation example, the reduced set derivation unit 44 adds only the prediction parameter whose appearance ratio is higher than a predetermined value to the reduced set RS in the set of prediction parameters decoded from the sub-blocks belonging to the first group. Therefore, the above-described problem that it may be difficult to effectively reduce the code amount when the number of sub-blocks is large can be solved.

  As described above, as described in generation examples 1 to 4 of the reduced set RS, the reduced set RS can be generated based on the prediction parameters belonging to the first group. More precisely, the reduced set RS can be generated based on at least one of the types of prediction parameters belonging to the first group or the appearance ratio of each prediction parameter belonging to the first group.

(Second prediction parameter decoding unit 45)
Next, the operation of the second prediction parameter decoding unit 45 will be described with reference to FIG. The second prediction parameter decoding unit 45 includes, among the encoded data for each subblock included in the subblock encoded data # 141b, each subblock determined by the group determination unit 41 as belonging to the second group. The prediction parameter P used for prediction is decoded.

  In other words, the second prediction parameter decoding unit 45 refers to the information related to the prediction parameter included in the sub-block encoded data # 141b and related to the prediction parameter for each sub-block belonging to the second group. Then, the prediction parameter P used for prediction of each sub-block belonging to the second group is decoded.

  The decoded prediction parameter P is output as prediction parameter # 45.

  FIG. 7 is a flowchart showing an example of the flow of decoding processing in the second prediction parameter decoding unit 45.

  As shown in FIG. 7, first, the second prediction parameter decoding unit 45 counts the number N of prediction parameters included in the reduced set RS (step S501).

  Subsequently, the second prediction parameter decoding unit 45 determines whether or not the number N of prediction parameters included in the reduced set RS is 1 (step S502).

  When N = 1 (Yes in step S502), the second prediction parameter decoding unit 45 sets the only prediction parameter included in the reduced set as the prediction parameter P (step S503).

  When N is not 1 (No in step S502), the second prediction parameter decoding unit 45 derives a prediction parameter estimated value Q (step S504). Here, the prediction parameter estimated value Q is a prediction parameter used for prediction of a subblock adjacent on the left or above the prediction target subblock. In addition, when a sub-block adjacent to the prediction target sub-block above or to the left is not decoded, it is a decoded sub-block above or to the left of the prediction target sub-block, and is the sub-block closest to the prediction target sub-block. The prediction parameter used for block prediction may be the estimated value Q.

  Subsequently, the second prediction parameter decoding unit 45 decodes a flag indicating whether or not the prediction parameter to be decoded is the same as the prediction parameter estimated value Q, and substitutes the decoded value into the variable a.

  Hereinafter, a case where the prediction parameter estimated value Q is the same as any prediction parameter included in the reduced set RS will be described as an example. In the following, the case where the value of the variable a is 1 corresponds to the case where the prediction parameter to be decoded is the same as the prediction parameter estimated value Q, and the case where the value of the variable a is not 1 The following description will be made assuming that the parameter is not the same as the predicted parameter estimated value Q.

  Subsequently, the second prediction parameter decoding unit 45 determines whether or not the value of the variable a is 1 (step S506).

  When the value of the variable a is 1 (Yes in step S506), the prediction parameter estimated value Q is set to the prediction parameter P (step S507).

  When the value of the variable a is not 1 (No in Step S506), the second prediction parameter decoding unit 45 determines whether or not the number N of prediction parameters included in the reduced set RS is 2 (Step S506). S508).

  When N = 2 (Yes in step S508), the second prediction parameter decoding unit 45 sets a prediction parameter that is included in the reduced set RS and does not match the prediction parameter estimated value Q as the prediction parameter P. (Step S509).

  When N is not 2 (No in step S508), the second prediction parameter decoding unit 45 decodes a bit string having a length of ceil (log2 (N-1)) bits and assigns the decoded value to the variable b ( Step S510). Here, ceil (...) Is a ceiling function whose value is the smallest integer among the integers greater than or equal to the value in parentheses (the same applies hereinafter). Therefore, ceil (...) Can also be expressed as a function that rounds up the value in the parentheses to an integer when the value in the parentheses is positive.

  For example, when N = 5, the second prediction parameter decoding unit 45 decodes a bit string having a length of ceil (log2 (5-1)) = 2 bits, and substitutes the decoded value into the variable b. Here, the value of the variable b takes one of the values b = 0, 1, 2, and 3 corresponding to the length of the bit string being 2 bits.

  Subsequently, the second prediction parameter decoding unit 45 calculates a prediction parameter having a (b + 1) th smallest index among prediction parameters that are included in the reduced set RS and do not match the prediction parameter estimation value Q. The prediction parameter P is set (step S511).

  For example, when the value of the variable b is 0, the second prediction parameter decoding unit 45 is the first of the prediction parameters that are included in the reduced set RS and that do not match the prediction parameter estimated value Q. A prediction parameter having a small index is set as the prediction parameter P.

  In addition, when the prediction parameter estimated value Q derived | led-out by the process demonstrated in step S504 is not the same as any prediction parameter included in the reduction set RS, it is included in the reduction set RS as the said prediction parameter estimation value Q. Of the prediction parameters, the prediction parameter with the smallest index may be used.

  The above is an example of the decoding process by the second prediction parameter decoding unit 45. A 2nd prediction parameter decoding part outputs the prediction parameter P decoded by the above processes as prediction parameter # 45.

  As described above, the video decoding device 1 uses the difference between the original image and the predicted image to determine which prediction parameter is selected from among a plurality of prediction parameters that specify the generation method of the predicted image for each prediction unit. An image decoding apparatus that decodes encoded data obtained by encoding together with selection information indicating a plurality of prediction units included in each of a plurality of unit regions constituting a prediction image, in the first group or Classification means (group determination unit 41) for classification into a second group is provided, and among the selection information, first selection information (subblock encoded data) that is selection information for each prediction unit belonging to the first group Reference is made to information relating to prediction parameters included in # 141b and relating to prediction parameters for each sub-block belonging to the first group) First selection means (first prediction parameter decoding unit 43) for selecting a prediction parameter for designating a prediction image generation method for each prediction unit belonging to the first group from a basic set of predetermined prediction parameters; , Out of the selection information, second selection information that is selection information for each prediction unit belonging to the second group (information related to a prediction parameter included in sub-block encoded data # 141b, With reference to the prediction parameter for each sub-block belonging to the group, a prediction parameter for specifying a prediction image generation method in each prediction unit belonging to the second group is selected as the first selection means (first prediction parameter). A reduced set including at least a part of the prediction parameters selected by the decoding unit 43), It is characterized in that it includes a second selecting means for selecting from the reduction set RS consisting prediction parameter less than or equal to the number of prediction parameters included in the basic set (second prediction parameter decoding section 45), the.

(Another configuration example of the prediction parameter decoding unit 144)
In the above description regarding the prediction parameter decoding unit 144, the configuration in which the reduced set deriving unit 44 generates the reduced set RS for each macroblock has been described, but the present invention is not limited to this.

  In the prediction parameter decoding unit 144, for example, the reduced set derivation unit 44 generates a reduced set RS for each sub-block, and the second prediction parameter decoding unit 45 based on the reduced set RS generated for each sub-block. The prediction parameter for the prediction target sub-block may be decoded.

  In such a configuration, the reduced set deriving unit 44 can generate the reduced set RS by performing the following processing, as shown in FIG.

(Step S701)
First, the reduced set derivation unit 44 initializes the reduced set RS by setting the reduced set RS to be empty.

(Step S702)
Subsequently, the reduced set deriving unit 44 sets a region composed of sub-blocks around the prediction target sub-block as the neighboring sub-block region NSR.

  FIG. 8B is a diagram illustrating an example of the neighboring sub-block region NSR. As shown in FIG. 8B, the neighboring sub-block region NSR is, for example, a sub-block around the prediction target sub-block, and the distance from the prediction target sub-block is 1 to 3 in units of sub-blocks. It can consist of sub-blocks that are within the city distance. Here, the urban area distance is a distance defined by the sum of absolute values of differences between coordinates with respect to coordinates between two points.

  Further, as shown in FIG. 8B, the neighboring sub-block region NSR may generally include a sub-block belonging to a macro block other than the macro block to which the prediction target sub-block belongs.

(Step S703)
Subsequently, the reduced set deriving unit 44 adds a decoded prediction parameter among the prediction parameters for each subblock included in the neighboring subblock region NSR to the reduced set RS.

  When the same prediction parameter corresponds to a plurality of subblocks included in the neighboring subblock region NSR, the reduced set derivation unit 44 sets the same parameter to 1 for the reduced set RS. Add only one.

  By performing the above operation, the reduced set deriving unit 44 can generate a reduced set RS for each macroblock. Moreover, the 2nd prediction parameter decoding part 45 can decode the prediction parameter with respect to a prediction object subblock based on the reduction | restoration set RS produced | generated for every subblock.

  In general, a prediction parameter for a prediction target sub-block has a correlation with a prediction parameter for sub-blocks around the prediction target sub-block. Therefore, the prediction parameter included in the reduced set RS generated by the above processing is highly likely to include the most appropriate prediction parameter in the prediction of the sub-blocks belonging to the second group. In addition, the number of prediction parameters included in the reduced set RS generated by the above processing is generally smaller than the number of prediction parameters that can be selected for the first group.

  Therefore, the moving image encoding apparatus that generates encoded data # 1 generates encoded data # 1 with a small code amount without sacrificing encoding efficiency by adopting a configuration corresponding to the above configuration. can do. Further, the moving picture decoding apparatus 1 can decode the encoded data # 1 generated in such a manner with a small code amount by adopting the above configuration.

  Note that when there is no decoded prediction parameter for a plurality of subblocks included in the neighboring subblock region NSR, the second prediction parameter decoding unit 45 selects, for example, a prediction parameter from the basic parameter set. What is necessary is just composition.

  Further, the reduced set deriving unit 44 in the present configuration example derives the reduced set RS by a process substantially similar to the process described in (Generated Example 1 of Reduced Set RS) to (Generated Example 4 of Reduced Set RS). It is good also as a simple structure. However, in that case, “first group” in (reduction set RS generation example 1) to (reduction set RS generation example 4) is read as “neighboring sub-block region NSR” in this configuration example.

  In the above description, the reduced set RS is used for the second group, but the present invention is not limited to this. The above-described processing can be applied to all subblocks in the macroblock. That is, it is good also as a structure which decodes a prediction parameter based on the reduction | decrease set RS produced | generated for every subblock with respect to all the subblocks in a macroblock.

  The moving picture encoding apparatus that generates the encoded data # 1 can further reduce the amount of prediction parameter codes for all the sub-blocks in the macroblock by adopting a configuration corresponding to the above configuration. Therefore, the moving picture encoding apparatus can generate encoded data # 1 with a smaller code amount. Moreover, the moving image decoding apparatus 1 can decode the encoded data # 1 generated as described above by adopting the above configuration.

(Predicted image generation unit 145)
Hereinafter, the generation process of the predicted image # 145 in the predicted image generation unit 145 will be described.

  The predicted image generation unit 145 determines the predicted pixel value of each pixel (prediction target pixel) in the predicted image # 145 (prediction target sub-block) according to the prediction direction (prediction mode) indicated by the prediction parameter # 144, for example, as follows: Generate as follows. In the following description, the case where the prediction parameter # 144 is any one of the prediction modes 0 to 8 shown in FIG. 5 will be described as an example.

  The prediction image generation unit 145 performs the following operation after assigning the prediction mode indicated by the prediction parameter # 144 to the prediction target pixel.

  When the allocated prediction mode is other than the prediction mode 2 (DC prediction), the predicted image generation unit 145 starts on the virtual line segment that faces the reverse direction of the prediction direction starting from the pixel position of the prediction target pixel. Among the decoded pixels that are positioned, the pixel value of the pixel closest to the pixel (hereinafter referred to as the nearest pixel) is set as the pixel value of the prediction target pixel. Further, the pixel value of the prediction target pixel may be a value calculated using the pixel value of the nearest pixel and the pixel values of the pixels around the nearest pixel.

  When the allocated prediction mode is prediction mode 2, the sub-block adjacent to the upper side of the prediction target sub-block (hereinafter referred to as the upper sub-block) and the sub-block adjacent to the left side (hereinafter referred to as the left sub-block) Is already decoded, the pixel value of the pixel in the lowermost column of the upper sub-block and the pixel value of the pixel in the rightmost column of the left sub-block are used as the pixel value of the prediction target pixel.

  If the assigned prediction mode is prediction mode 2 and the upper sub-block has been decoded and the left sub-block has not been decoded, the pixel value of the pixel in the bottom row of the upper sub-block, and , The average value of the pixel values of the pixels in the rightmost column in the subblock to the left of the prediction target subblock and closest to the prediction target subblock (hereinafter referred to as the left nearest subblock), The pixel value of the prediction target pixel is used.

  If the assigned prediction mode is prediction mode 2 and the upper sub-block has not been decoded and the left sub-block has been decoded, the sub-block above the prediction target sub-block is The average value of the pixel value of the pixel in the lowermost column in the subblock closest to the prediction target subblock (hereinafter referred to as the upper nearest subblock) and the pixel value of the pixel in the rightmost column of the left subblock, The pixel value of the prediction target pixel is used.

  If the assigned prediction mode is prediction mode 2 and neither the upper subblock nor the left subblock has been decoded, the pixel value of the pixel in the bottom row of the uppermost subblock, The average value of the pixel values of the rightmost column of the left nearest subblock is set as the pixel value of the prediction target pixel.

  Hereinafter, an example of the generation process of the predicted image # 145 by the predicted image generation unit 145 when the prediction target sub-block is 4 × 4 pixels will be described more specifically with reference to FIG.

  FIG. 9 is a diagram illustrating each pixel (prediction target pixel) of a prediction target sub-block that is 4 × 4 pixels and pixels (reference pixels) around the prediction target sub-block. As illustrated in FIG. 9, the prediction target pixels are denoted by reference signs a to p, the reference pixels are denoted by reference signs A to M, and the pixel value of the pixel X (X is any one of a to p and A to M) is set to X. It will be expressed as Further, it is assumed that the reference pixels A to M have all been decoded.

(Prediction mode 0)
When the allocated prediction mode is the prediction mode 0, the predicted image generation unit 145 converts the pixel values a to p into the following formulas a, e, i, m = A,
b, f, j, n = B,
c, g, k, o = C,
d, h, l, p = D
Generate by.

(Prediction mode 2)
When the allocated prediction mode is the prediction mode 2 (DC prediction), the prediction image generation unit 145 converts the pixel values a to p into the following expressions: a to p = ave (A, B, C, D, I, J , K, L)
Generate by. Here, ave (...) Indicates that an element included in parentheses is averaged.

(Prediction mode 4)
When the allocated prediction mode is the prediction mode 4, the predicted image generation unit 145 converts the pixel values a to p into the following formula d = (B + (C × 2) + D + 2) >> 2,
c, h = (A + (B × 2) + C + 2) >> 2,
b, g, l = (M + (A × 2) + B + 2) >> 2,
a, f, k, p = (I + (M × 2) + A + 2) >> 2,
e, j, o = (J + (I × 2) + M + 2) >> 2,
i, n = (K + (J × 2) + I + 2) >> 2,
m = (L + (K × 2) + J + 2) >> 2
Generate by. Here, “>>” represents a right shift operation, and for any positive integer x, s, the value of x >> s is equal to the value obtained by rounding down the decimal part of x ÷ (2 ＾ s).

  Further, the predicted image generation unit 145 can calculate the pixel values a to p by the same method for prediction modes other than the prediction mode.

<Additional notes regarding video decoding device>
Although the video decoding device according to the present invention has been described above, the present invention is not limited to the above configuration.

(Appendix 1)
For example, the second prediction parameter decoding unit 45 may be configured to switch whether to use the reduced set RS when decoding the prediction parameter # 45 according to the flag included in the encoded data # 1. .

  More specifically, for example, when the value of the flag included in the encoded data # 1 is 1, the second prediction parameter decoding unit 45 decodes the prediction parameter # 45 using the reduced set RS, When the value of the flag included in the encoded data # 1 is 0, the prediction parameter # 45 may be decoded using the basic parameter set instead of the reduced set RS.

  With this configuration, it is possible to reduce the processing amount when decoding the prediction parameter # 45.

(Appendix 2)
Further, the second prediction parameter decoding unit 45 may be configured to switch whether to use the reduced set RS when decoding the prediction parameter # 45 according to the number of subblocks included in the macroblock. .

  More specifically, when the number of subblocks included in the macroblock is 16 or more, the prediction parameter # 45 is decoded using the reduced set RS, and the number of subblocks included in the macroblock is 16. If the number is less than the number, the prediction parameter # 45 may be decoded using the basic parameter set instead of the reduced set RS.

  With this configuration, it is possible to reduce the processing amount when decoding the prediction parameter # 45.

(Appendix 3)
In the above description, a prediction parameter set as shown in FIG. 5 is used as the basic parameter set, but the present invention is not limited to this.

  For example, the prediction parameter decoding unit 144 sets, as the basic parameter set, a parameter set that emphasizes the horizontal direction as shown in FIG. 16A or a parameter that emphasizes the vertical direction as shown in FIG. It is good also as a structure which uses a set.

  More specifically, for example, when the edge in the horizontal direction exists in the macroblock, the prediction parameter decoding unit 144 sets the horizontal direction as shown in FIG. When an important parameter set is used and a vertical edge exists in the macroblock, a parameter set that emphasizes the vertical direction as shown in FIG. 16B is used as the basic parameter set. It can be configured.

  In addition, when such a plurality of basic parameter sets are selectively used and the parameter set shown in FIG. 16A or 16B is selected as the basic parameter set, the second prediction parameter decoding is performed. The unit 45 decodes the prediction parameter # 45 using the reduced set RS, and when the parameter set shown in FIG. 5 is selected as the basic parameter set, the second prediction parameter decoding unit 45 replaces the reduced set RS. Thus, the prediction parameter # 45 may be decoded using the basic parameter set.

  With such a configuration, the reduced set RS can be used according to the characteristics of the image in the macroblock.

(Moving picture encoding device)
Below, the moving image encoder (image encoder) 2 which concerns on this embodiment is demonstrated with reference to FIGS. FIG. 10 is a block diagram illustrating a configuration of the moving image encoding device 2. As shown in FIG. 10, the moving picture encoding apparatus 2 includes a header information determination unit 21, a header information encoding unit 22, an MB setting unit 23, an MB encoding unit 24, a variable length code multiplexing unit 25, and an MB decoding unit. 26 and a frame memory 27.

  Generally speaking, the moving image encoding device 2 is a device that generates and outputs encoded data # 1 by encoding the input image # 100.

  The header information determination unit 21 determines header information based on the input image # 100. The determined header information is output as header information # 21. The header information # 21 includes the image size of the input image # 100. The header information # 21 is input to the MB setting unit 23 and supplied to the header information encoding unit 22.

  The header information encoding unit 22 encodes header information # 21 and outputs encoded header information # 22. The encoded header information # 22 is supplied to the variable length code multiplexer 25.

  The MB setting unit 23 divides the input image # 100 into a plurality of macro blocks based on the header information # 21, and outputs a macro block image # 23 related to each macro block. The macro block image # 23 is sequentially supplied to the MB encoding unit 24.

  The MB encoding unit 24 encodes sequentially input macroblock image # 23 to generate MB encoded data # 24. The generated MB encoded data # 24 is supplied to the variable length code multiplexer 25. Since the configuration of the MB encoding unit 24 will be described later, the description thereof is omitted here.

  The variable length code multiplexing unit 25 generates encoded data # 1 by multiplexing encoded header information # 22 and MB encoded data # 24, and outputs the encoded data # 1.

  The MB decoding unit 26 sequentially decodes the MB encoded data # 24 corresponding to each input macroblock, thereby generating and outputting a decoded image # 26 corresponding to each macroblock. The decoded image # 26 is supplied to the frame memory 27.

  In the frame memory 27, the input decoded image # 26 is recorded. At the time of encoding a specific macroblock, decoded images corresponding to all macroblocks that precede the macroblock in the raster scan order are recorded in the frame memory.

(MB encoding unit 24)
Hereinafter, the MB encoding unit 24 will be described more specifically with reference to different drawings.

  FIG. 11 is a block diagram showing a configuration of the MB encoding unit 24. As shown in FIG. 11, the MB encoding unit 24 includes a sub-block dividing unit 241, a prediction parameter determining unit 242, a prediction parameter encoding unit 243, a prediction residual generating unit 244, a transform coefficient encoding unit 245, a prediction residual. A decoding unit 246, a sub-block decoded image generation unit 247, a predicted image generation unit 248, and an MB encoded data generation unit 249 are provided.

  The subblock division unit 241 divides the macroblock image # 23 into a plurality of subblocks, subblock position information # 241a indicating the position of each subblock constituting the macroblock in the macroblock, and the subblock Sub-block image # 241b, which is image data related to the sub-block indicated by position information # 241a, is sequentially output in a predetermined order.

  The sub-block dividing unit 241 outputs sub-block position information # 241a and sub-block image # 241b related to sub-blocks belonging to the first group described later, and then outputs sub-blocks related to sub-blocks belonging to the second group described later. It is preferable that the position information # 241a and the sub-block image # 241b are output. For example, it is preferable that the sub blocks belonging to the first group are scanned in the raster scan order, and then the sub blocks belonging to the second group are scanned in the raster scan order.

  The prediction parameter determination unit 242 determines and outputs a prediction parameter # 242 used for generating a prediction image related to the subblock indicated by the subblock position information # 241a. Further, the prediction parameter encoding unit 243 encodes the prediction parameter # 242 and outputs the encoded prediction parameter # 243. Since the configuration of the prediction parameter determination unit 242 and the prediction parameter encoding unit 243 will be described later, description thereof is omitted here.

  The prediction residual generation unit 244 identifies a subblock to be predicted based on the subblock position information # 241a, and the prediction generated by the subblock image # 241b and the prediction image generation unit 248 in the subblock. A prediction residual # 244 that is a difference image from the image # 248 is generated.

  The transform coefficient coding unit 245 generates a transform coefficient of the prediction residual # 244 by applying frequency transform having the same size as the subblock size to the prediction residual # 244.

  Further, the transform coefficient encoding unit 245 quantizes the transform coefficient to generate a quantized transform coefficient # 245a, and then applies a variable length coding method such as CABAC or CAVLC to the quantized transform coefficient # 245a. This is applied to generate a variable length code, and the variable length code is output as encoded data # 245b.

  The prediction residual decoding unit 246 generates and outputs a decoding residual # 246 by inversely quantizing the quantized transform coefficient # 245a and then applying inverse transform of frequency transform (inverse frequency transform).

  Note that the above processing by the prediction residual generation unit 244, the transform coefficient encoding unit 245, and the prediction residual decoding unit 246 does not limit the present invention. For example, the transform coefficient coding unit 245 may omit the frequency transform and directly quantize the prediction residual.

  The sub-block decoded image generation unit 247 generates and outputs a sub-block decoded image # 247 by adding the predicted image # 248 and the decoded residual # 246.

  The predicted image generation unit 248 generates and outputs a predicted image # 248 corresponding to the prediction target sub-block based on the prediction parameter # 242, the decoded image # 27, and the sub-block decoded image # 247. As a specific generation method of the prediction image # 248 in the prediction image generation unit 248, for example, a method similar to the generation method of the prediction image # 145 in the prediction image generation unit 145 described above can be applied.

  The MB encoded data generation unit 249 accumulates the encoded data # 245b for each subblock and the encoded prediction parameter # 243 for each subblock, and integrates the encoded data in units of macroblocks. MB encoded data # 24, which is data, is generated and output.

  Hereinafter, the prediction parameter determination unit 242 and the prediction parameter encoding unit 243 will be described with reference to different drawings.

(Prediction parameter determination unit 242)
FIG. 12 is a block diagram illustrating a configuration of the prediction parameter determination unit 242. As illustrated in FIG. 12, the prediction parameter determination unit 242 includes a group determination unit 51, a switch unit 52, a first prediction parameter determination unit 53, a reduced set derivation unit 54, and a second prediction parameter determination unit 55. .

  The group determination unit 51 classifies the sub-block indicated by the sub-block position information # 241a into one of a plurality of groups, and outputs group information # 51 indicating the classification result to the switch unit 52.

  The group determination unit 51 can classify each sub-block into either the first group or the second group, for example, as already described with reference to FIGS.

  Moreover, the group determination part 51 may classify | categorize each subblock in any group among several groups using the classification method different for every macroblock. For example, the sub-blocks constituting the macro block MB1 are classified into two groups as shown in FIGS. 4A to 4B, and the sub-blocks constituting the macro block MB2 different from the macro block MB1 are classified. As shown in (c) to (d) of FIG. 4, it may be configured to be classified into two groups. As described above, when any one of a plurality of classification methods is used for each macroblock, the group determination unit 51 may output a flag indicating which classification method is used. preferable. By transmitting the flag to the moving picture decoding apparatus that decodes the encoded data # 1, the moving picture decoding apparatus can identify which classification method is used in the group determination unit 51.

  Based on the group information # 51, the switch unit 52 converts the subblock encoded data # 241b, which is encoded data related to the subblock indicated by the subblock position information # 241a, to the first prediction parameter determination unit 53 or the second This is transmitted to one of the prediction parameter determination units 55.

  Specifically, in the group determination unit 51, when the subblock indicated by the subblock position information # 241a is classified into the first group, the switch unit 52 converts the subblock encoded data # 241b into the first block. 1 is transmitted to the prediction parameter determination unit 53. When the group determination unit 51 determines that the sub-block indicated by the sub-block position information # 241a is classified into the second group, the sub-block encoded data # 241b Is transmitted to the second prediction parameter determination unit 55.

  The first prediction parameter determination unit 53 is used to generate a prediction image for each subblock belonging to the first group based on the decoded image # 27, the subblock decoded image # 247, and the subblock encoded data # 241b. The prediction parameter # 53 is determined (selected) and output. The prediction parameter # 53 is also supplied to the reduced set derivation unit 54.

  For example, if the prediction parameter is H.264. In the case of the intra prediction mode in the H.264 / MPEG-4 AVC standard, the first prediction parameter determination unit 53 applies the basic parameter set shown in FIG. 5 described above to each sub-block belonging to the first group. One of the prediction modes is selected and output.

  Note that the specific method for determining the prediction parameter # 53 in the first prediction parameter determination unit 53 does not limit the present invention. For example, the first prediction parameter determination unit 53 includes each sub-group belonging to the first group. The prediction parameter # 53 may be determined for the block so that the difference between the predicted image and the input image # 100 in the sub-block is minimized. For example, the first prediction parameter determination unit 53 has the largest difference between the prediction image generated using the prediction mode 1 and the input image # 100 in the basic parameter set for the sub-block SB1 belonging to the first group. In such a case, the prediction mode 1 is selected for the sub-block SB1, and the sub-block SB2 belonging to the first group is generated using the prediction mode 2 in the basic parameter set. In the case where the difference between the predicted image and the input image # 100 is minimized, the prediction mode 2 may be selected for the sub-block SB2.

  (A) of FIG. 13 is a figure which shows the example of the prediction mode which the 1st prediction parameter determination part 53 selected with respect to each subblock which belongs to a 1st group among each subblock which comprises macroblock MB. is there. In the example shown in (a) of FIG. 13, the first prediction parameter determination unit 53 determines which one of prediction mode 1, prediction mode 6, and prediction mode 8 for each sub-block belonging to the first group. The prediction mode is selected. In this example, the prediction mode 1, the prediction mode 6, and the prediction mode 8 are supplied to the reduced set derivation unit 54 as the prediction parameter # 53.

  Through the above operation, the reduced set derivation unit 54 is supplied with the prediction parameter # 53 related to each sub-block belonging to the first group.

  The configuration of the reduced set derivation unit 54 is the same as that of the reduced set derivation unit 44 already described. That is, the reduced set derivation unit 54 generates a reduced set RS using the prediction parameter # 53. The generation method of the reduced set RS in the reduced set derivation unit 54 is the same as the generation method of the reduced set RS in the reduced set derivation unit 44 already described.

  When a plurality of generation methods are selectively used in the reduced set derivation unit 54, the reduced set derivation unit 54 preferably outputs a flag indicating which generation method has been selected. By transmitting the flag to the video decoding device that decodes the encoded data # 1, the video decoding device can identify which generation method is used in the reduced set deriving unit 54.

  FIG. 13B is a diagram illustrating an example of a reduced set RS generated by the reduced set derivation unit 54 when each prediction mode illustrated in FIG. 13A is supplied as the prediction parameter # 53. . As shown in (b) of FIG. 13, in this example, the reduced set RS includes a prediction mode 1, a prediction mode 8, and a prediction mode 6.

  The second prediction parameter determination unit 55 selects and outputs a prediction parameter # 55 used for generating a prediction image for each sub-block belonging to the second group from the prediction parameters included in the reduced set RS.

  The specific method for determining the prediction parameter # 55 in the second prediction parameter determination unit 55 is not limited to the present invention. For example, the second prediction parameter determination unit 55 applies to each sub-block belonging to the second group. On the other hand, the prediction parameter # 55 that can most appropriately generate the prediction image in the sub-block may be selected from the prediction parameters included in the reduced set RS.

  FIG. 13C shows the second prediction parameter determination unit 55 shown in FIG. 13B for each subblock belonging to the second group among the subblocks constituting the macroblock MB. It is a figure which shows the example of the prediction mode selected from reduction set RS. As shown in FIG. 13C, for each sub-block belonging to the second block, prediction mode 1, prediction mode 6 or prediction mode included in the reduced set RS shown in FIG. 13B. Any one of 8 prediction modes is selected.

  In general, there is a correlation between optimal prediction parameters for each sub-block constituting a macroblock. Therefore, the prediction parameter selected for each sub-block belonging to the first group is highly likely to be an optimal prediction parameter for each sub-block belonging to the second group.

  In addition, as described above, the second prediction parameter determination unit 55 selects the prediction parameter # 55 for each sub-block belonging to the second group from the prediction parameters included in the reduced set RS, so that the reduced set RS is not used. As compared with the above, the code amount of the prediction parameter # 55 can be reduced.

  For example, as will be described later, when the prediction mode is encoded together with a flag indicating whether or not the estimated value is the same, the three prediction modes included in the reduced set RS shown in FIG. (Log2 (3-1)) = 1 can be encoded using a code amount of 1 bit. On the other hand, when the second prediction parameter determination unit 55 selects a prediction parameter from a basic parameter set composed of nine prediction modes without using the reduced set RS, ceil (log2 (9-1)) = A code amount of 3 bits is required. Further, in the above example, since the second group is composed of 8 sub-blocks, 3 × for each macroblock is obtained by using the reduced set RS as compared with the case where the reduced set RS is not used. The code amount of 8-1 × 8 = 16 bits can be reduced.

  In general, when a prediction parameter is encoded together with a flag indicating whether or not it is the same as the estimated value, the number of prediction parameters that can be selected for each sub-block belonging to the first group is included in Nfs and the reduced set RS. Assuming that the number of prediction parameters is Nrs and the number of sub-blocks included in the second group is Ngs, the use of the reduced set RS makes it possible to obtain Ngs for each macroblock as compared with the case where the reduced set RS is not used. The code amount of x (ceil (log2 (Nfs-1))-ceil (log2 (Nrs-1))) bits can be reduced.

  Thus, by using the reduced set RS, it is possible to reduce the amount of code necessary for encoding the prediction parameter without sacrificing the encoding efficiency.

(Another configuration example of the prediction parameter determination unit 242)
In the above description regarding the prediction parameter determination unit 242, the configuration in which the reduced set derivation unit 54 generates the reduced set RS for each macroblock has been described, but the present invention is not limited to this.

  That is, in the prediction parameter determination unit 242, the reduced set derivation unit 54 generates a reduced set RS for each sub-block, and the second prediction parameter determination unit 55 based on the reduced set RS generated for each sub-block. The prediction parameter for the prediction target sub-block may be determined.

  In such a configuration, the reduced set deriving unit 54 operates in the same manner as the operation of the reduced set deriving unit 44 described in (Step S701) to (Step S703) of (another configuration example of the prediction parameter decoding unit 144). The configuration may be such that However, the decoded prediction parameters in (step S701) to (step S703) of (another configuration example of the prediction parameter decoding unit 144) correspond to the encoded prediction parameters in this example.

  Thereby, the reduced set deriving unit 54 can generate a reduced set RS for each sub-block. Moreover, the 2nd prediction parameter determination part 55 can determine the prediction parameter with respect to a prediction object subblock based on the reduction | restoration set RS produced | generated for every subblock.

  In general, a prediction parameter for a prediction target sub-block has a correlation with a prediction parameter for sub-blocks around the prediction target sub-block. Therefore, the prediction parameter included in the reduced set RS generated by the above processing is highly likely to include the most appropriate prediction parameter in the prediction of the sub-blocks belonging to the second group. In addition, the number of prediction parameters included in the reduced set RS generated by the above processing is generally smaller than the number of prediction parameters that can be selected for the first group.

  Therefore, the moving image encoding apparatus 1 can generate encoded data # 1 with a small code amount without sacrificing encoding efficiency by adopting the above configuration.

  Note that when there is no encoded prediction parameter for a plurality of subblocks included in the neighboring subblock region NSR, the second prediction parameter determination unit 55 selects, for example, a prediction parameter from the basic parameter set. What is necessary is just to make it a structure.

  Further, the reduced set derivation unit 54 in the present configuration example derives the reduced set RS by a process substantially similar to the process described in (Generation example 1 of the reduced set RS) to (Generation example 4 of the reduced set RS). It is good also as a simple structure. However, in that case, “first group” in (reduction set RS generation example 1) to (reduction set RS generation example 4) is read as “neighboring sub-block region NSR” in this configuration example.

  In the above description, the reduced set RS is used for the second group, but the present invention is not limited to this. The above-described processing can be applied to all subblocks in the macroblock. That is, it is good also as a structure which determines a prediction parameter based on the reduction | restoration set RS produced | generated for every subblock with respect to all the subblocks in a macroblock.

  By adopting the above configuration, it is possible to reduce the amount of prediction parameter codes for all sub-blocks in a macroblock. Therefore, the moving image encoding apparatus 1 can generate encoded data # 1 with a smaller code amount by adopting the above configuration.

(Prediction parameter encoding unit 243)
Next, the prediction parameter encoding unit 243 will be described with reference to FIG. FIG. 14 is a block diagram illustrating a configuration of the prediction parameter encoding unit 243. As illustrated in FIG. 14, the prediction parameter encoding unit 243 includes a group determination unit 61, a switch unit 62, a first prediction parameter encoding unit 63, a reduced set derivation unit 64, and a second prediction parameter encoding unit 65. I have.

  The group determination unit 61 has substantially the same configuration as the group determination unit 51 already described. That is, the group determination unit 61 classifies the sub-block indicated by the sub-block position information # 241a into one of a plurality of predetermined groups, and sets the group information # 61 indicating the classification result to the switch unit 62. Output. Note that the group determination unit 61 may classify each sub-block into one of a plurality of predetermined groups using the same classification method as that used in the group determination unit 51.

  Based on the group information # 61, the switch unit 62 sets the prediction parameter # 242 related to the subblock indicated by the subblock position information # 241a to the first prediction parameter encoding unit 63 or the second prediction parameter encoding unit 65. Of these, it is transmitted to one of the parameter encoding units.

  Specifically, when the group determination unit 61 classifies the subblock indicated by the subblock position information # 241a into the first group, the switch unit 62 sets the prediction parameter # 242 as the first prediction parameter. When the subblock indicated by the subblock position information # 241a is classified into the second group in the group determination unit 61, the prediction parameter # 242 is transmitted to the encoding unit 63 and the reduced set derivation unit 64. Is transmitted to the second prediction parameter encoding unit 65.

  The 1st prediction parameter encoding part 63 produces | generates and outputs the encoding prediction parameter # 63 by encoding prediction parameter # 242 regarding each subblock which belongs to a 1st group.

  Specifically, the first prediction parameter encoding unit 63 first sets the prediction parameter selected for the subblocks around each subblock belonging to the first group to the estimated value for the subblock.

  Subsequently, the first prediction parameter encoding unit 63 encodes a flag indicating whether or not the prediction parameter selected for the subblock is different from the estimated value, and is further selected for the subblock. If the predicted parameter is different from the estimated value, the predicted parameter is encoded.

  Here, when the prediction parameter of each sub-block is selected from the basic parameter set, the prediction parameter can be expressed by a 1-bit or 4-bit code including a flag.

  As described above, by performing the encoding using the estimated value, it is possible to increase the compression rate when encoding the prediction parameter # 242 regarding each sub-block belonging to the first group.

  The first prediction parameter encoding unit 63 may be configured to encode the prediction parameter # 242 regarding each subblock belonging to the first group as it is.

  The reduced set derivation unit 64 generates a reduced set RS using the prediction parameter # 242 related to the sub-blocks belonging to the first group. The generation method of the reduced set RS in the reduced set derivation unit 64 is the same as the generation method of the reduced set RS in the reduced set derivation unit 44 described above, and thus the description thereof is omitted here.

  The second prediction parameter encoding unit 65 encodes a prediction parameter included in the reduced set RS, that is, a prediction parameter selected for each sub-block belonging to the second group, thereby encoding encoded parameter # 65. Is generated and output.

  Specifically, the second prediction parameter encoding unit 65 first sets the prediction parameter selected for the subblocks around each subblock belonging to the second group to the estimated value for the subblock.

  When the number of prediction parameters Nrs included in the reduced set RS is 1, the encoding process of the second prediction parameter is terminated without encoding anything.

  Subsequently, the second prediction parameter encoding unit 65 encodes a flag indicating whether or not the prediction parameter selected for the subblock is different from the estimated value, and further selects the flag for the subblock. If the predicted parameter is different from the estimated value, the predicted parameter is encoded.

  Here, when Nrs is 2, the encoding process of the second prediction parameter ends.

  In other cases, the prediction parameter included in the reduced set can be expressed by a code of ceil (log2 (Nrs-1)) bits.

  Also, the number Nrs of prediction parameters included in the reduced set RS is generally smaller than the number of prediction parameters that can be selected for the first group.

  Therefore, by using the reduced set RS, the prediction parameter selected for each sub-block belonging to the second group can be encoded with a smaller code amount.

  Further, as described above, by performing the encoding using the estimated value, it is possible to increase the compression rate when encoding the prediction parameter # 242 regarding each sub-block belonging to the second group.

  Note that the second prediction parameter encoding unit 65 may be configured to encode the prediction parameters selected for each sub-block belonging to the second group as they are.

<Additional Notes on Video Encoding Device>
The video encoding device 2 according to the present invention has been described above, but the present invention is not limited to the above configuration.

(Appendix 1 ')
For example, the second prediction parameter determination unit 55 determines whether or not to use the reduced set RS according to the magnitude of the spatial correlation in the macroblock of the prediction parameter, and when the spatial correlation is small, the reduced set It is good also as a structure which determines prediction parameter # 55, without utilizing RS. In addition, when the second prediction parameter determination unit 55 determines the prediction parameter # 55 without using the reduced set RS, the moving image encoding device 2 encodes a flag indicating that the reduced set RS is not used. It is preferable to adopt a configuration for transmitting to the video decoding device.

  By adopting such a configuration, when the spatial correlation of the prediction parameter is small, the prediction parameter # 55 can be determined without using the reduced set RS, so that the processing amount for determining the prediction parameter is suppressed. be able to.

(Appendix 2 ')
The second prediction parameter determination unit 55 may be configured to switch whether to use the reduced set RS when determining the prediction parameter # 55 according to the number of subblocks included in the macroblock. .

  More specifically, when the number of subblocks included in the macroblock is 16 or more, the prediction parameter # 55 is determined using the reduced set RS, and the number of subblocks included in the macroblock is 16. When the number is less than the number, the prediction parameter # 45 may be determined using the basic parameter set instead of the reduced set RS.

  By adopting such a configuration, when the spatial correlation of the prediction parameter is small, the prediction parameter # 55 can be determined without using the reduced set RS, so that the processing amount for determining the prediction parameter is suppressed. Can do.

(Appendix 3 ')
Further, the prediction parameter determination unit 242 sets a basic parameter set as shown in FIG. 16A as a basic parameter set according to the characteristics of the image in the macroblock, or FIG. 16B. It is also possible to use a parameter set that emphasizes the vertical direction as shown in FIG.

  For example, the prediction parameter determination unit 242 uses a parameter set that emphasizes the horizontal direction as shown in FIG. 16A as the basic parameter set when there is a horizontal edge in the macroblock. When there is an edge in the vertical direction in the macroblock, it is possible to use a parameter set that emphasizes the vertical direction as shown in FIG. 16B as the basic parameter set. It is.

  In addition, when a plurality of basic parameter sets are selectively used and the parameter set shown in FIG. 16A or 16B is selected as the basic parameter set, the second prediction parameter determining unit 55 When the prediction parameter # 55 is determined using the reduced set RS and the parameter set shown in FIG. 5 is selected as the basic parameter set, the second prediction parameter determining unit 55 replaces the reduced set RS with The prediction parameter # 55 may be decoded using the basic parameter set.

  With such a configuration, the reduced set RS can be used according to the characteristics of the image in the macroblock, so that the code amount of the prediction parameter can be effectively reduced.

(Data structure of encoded data # 1)
Hereinafter, the data structure of the encoded data # 1 generated by the video encoding device 2 will be described with reference to FIG.

  FIG. 15 is a diagram illustrating a bit stream structure of bit stream #MBS for each macroblock of encoded data # 1. As shown in FIG. 15, the bitstream #MBS is subblock information # SB1 to #SBN that is information about subblocks SB1 to SBN (where N is the number of subblocks in the macroblock) included in the macroblock. (Where N is the number of sub-blocks in the macroblock).

  Further, as shown in FIG. 15, each sub-block information #SBn (1 ≦ n ≦ N) corresponds to sub-block position information #Ln, which is information indicating the position of sub-block SBn in the macroblock, and sub-block SBn. And prediction parameter information #Pn indicating the attached prediction parameter.

  The sub-block position information #Ln is information that is referred to in order to identify the position of the sub-block SBn in the macroblock in the video decoding device that decodes the encoded data # 1. In particular, in the video decoding device 1 described above, the sub-block position information #Ln is information that is referred to in order to classify the sub-blocks SBn into groups.

  Prediction parameter information #Pn is information for specifying a prediction parameter associated with sub-block SBn in a video decoding device that decodes encoded data # 1. In particular, in the video decoding device 1 described above, the prediction parameter information #Pn is information indicating one of the prediction parameters that can be selected for the group to which the sub-block SBn belongs.

  For example, when the sub-block SBn belongs to the first group and the prediction parameter set composed of prediction parameters selectable for the first group is a basic parameter set, the prediction parameter information # Pn is information indicating one of the prediction modes 0 to 8 included in the basic parameter set.

  Further, when the sub-block SBn belongs to the second group, the prediction parameter information #Pn is information indicating any prediction parameter among the prediction parameters included in the reduced set RS.

  When the encoded data # 1 is data generated by encoding a prediction mode together with a flag indicating whether or not the estimated value is the same as the estimated value, the prediction parameter information #Pn is ceil (log2 (N-1)) represented by a bit sign. Here, N is the number of prediction parameters that can be selected for the group to which the sub-block SBn belongs.

  In general, the number of prediction parameters included in the reduced set RS is smaller than the number of prediction parameters that can be selected for the first group. Therefore, the code amount of the prediction parameter included in the encoded data # 1 is small compared to the case where the reduced set RS is not used.

  Specifically, when the encoded data # 1 is data generated by encoding the prediction mode together with a flag indicating whether or not the estimated value is the same as the estimated value, the code of the prediction parameter information #Pn The amount is less ceil (log2 (Nfs-1))-ceil (log2 (Nrs-1)) bits than when the reduced set RS is not used. Here, Nfs is the number of prediction parameters that can be selected for each sub-block belonging to the first group, and Nrs is the number of prediction parameters included in the reduced set RS.

  Thus, the code amount of the encoded data # 1 can be reduced by using the reduced set RS.

  The encoded data # 1 may not include the sub-block position information #Ln. For example, if a common sub-block scan order is set in advance at the time of encoding and decoding, the position of the sub-block can be determined based on information about what number of sub-blocks in the encoded data # 1.

<Application example to other prediction parameters>
In the above description, description has been given mainly using the prediction mode in intra prediction as an example of the prediction parameter. However, the present invention is not limited to this, and in general, encoding / decoding processing of moving images The present invention can also be applied to other parameters used when generating a predicted image.

  Hereinafter, an application example when the prediction parameter is a motion vector in motion compensation prediction and an application example when the prediction parameter is a weight coefficient in luminance compensation prediction will be described.

(Application example to motion vectors)
In motion compensation prediction, the position of a region on a decoded image used for prediction of a prediction target sub-block is expressed using a prediction parameter called a motion vector.

  The motion vector is selected from a prediction parameter set whose number of elements depends on the image size and accuracy (interpolation accuracy). For example, when the image width (the number of pixels in the horizontal direction) is W, the height (the number of pixels in the vertical direction) is H, and the interpolation accuracy is 0.25 pixels, the motion vector V is defined by the following equation: Prediction parameter set S.

S≡ {V | V = ((N / 4), (M / 4))
Here, N and M are integers satisfying 0 ≦ N <4W and 0 ≦ M <4H.

  If the prediction parameter is such a motion vector, the first prediction parameter determination unit 53 may determine and assign a motion vector for each sub-block belonging to the first group, for example. Further, the first prediction parameter decoding unit 43 can decode a motion vector for each sub-block belonging to the first group.

  The reduced set derivation units 44 and 54 can generate a reduced set RS from the motion vectors allocated to the sub-blocks belonging to the first group. Further, as already described, the reduced set deriving units 44 and 54 may generate the reduced set RS from the motion vectors assigned to the subblocks around the prediction target subblock.

  With respect to the other units included in the video decoding device 1 and the video encoding device 2, the same operation as when the prediction parameter is in the prediction mode is performed, so that the prediction parameter included in the encoded data # 1 The amount of codes can be reduced.

  When the prediction parameter is a motion vector, the reduced set derivation units 44 and 54, for example, set the norm of the difference vector from the motion vector allocated to the subblocks around the prediction target subblock to be equal to or less than a certain value. Such a motion vector may be added to the reduced set RS.

  In general, when the prediction parameter is a motion vector, the prediction parameter set includes a large number of prediction vectors. For example, the number of motion vectors included in the prediction parameter set S is 8000 × 4000 when W = 2000 and H = 1000.

  On the other hand, the number of motion vectors stored in the reduced set RS is limited to the number of subblocks belonging to the first group or the number of subblocks around the prediction target subblock. For example, even when subblocks included in the neighboring subblock region NSR shown in FIG. 8B are used, the number of motion vectors stored in the reduced set RS is at most 24.

  Therefore, depending on the characteristics of the image to be encoded / decoded, even if the reduced set RS is generated only from the motion vectors assigned to the subblocks around the prediction target subblock, the reduced set RS is sufficient. Since a number of motion vectors are not accumulated, the prediction accuracy may be reduced and the code amount of residual data may be increased. The code amount of the residual data is increased by adding to the reduced set RS a motion vector such that the norm of the difference vector with respect to the motion vector assigned to the sub-blocks around the prediction target sub-block is equal to or less than a certain value. Therefore, it is possible to reduce the code amount of the prediction parameter.

(Application example to weighting coefficient in luminance compensation prediction)
In luminance compensation prediction, the luminance of a prediction target sub-block is predicted using a value obtained by multiplying the luminance of each of a plurality of reference pictures referenced by the prediction target sub-block for motion compensation prediction by a weighting coefficient.

  The present invention can also be applied to the case where the prediction parameter is the weighting factor. For example, the reduced set derivation units 44 and 54 may be configured to generate the reduced set RS from the weighting factors assigned to the subblocks around the prediction target subblock.

  When the present invention is applied to the weighting factor, each value taken by the weighting factor is associated with a plurality of representative values whose total number is determined in advance, and the representative value is used as a prediction parameter. It is good also as a structure.

  For example, when the value W of the weighting factor can take all real values (or values with a predetermined number of digits) satisfying 0 ≦ W ≦ 1, (n / X) ≦ Wn ≦ ((n + 1) ) / X) all weight coefficient values Wn may be associated with the representative value w, and the representative value w may be used as a prediction parameter. Here, X is a natural number, and n is an integer that satisfies 0 ≦ n ≦ X−1. By performing such association, all real values satisfying 0 ≦ W ≦ 1 can be mapped to a set of representative values Wn whose total number is X. Moreover, what is necessary is just to set the specific value of X so that the code amount of coding data may become smaller, for example.

  As described above, by limiting the number of elements used as the prediction parameter to a predetermined number, it is possible to reduce the code amount of the encoded data as compared with the case where the weighting factor is directly used as the prediction parameter.

(Additional notes)
An image encoding device according to the present invention, in an image encoding device that encodes a difference between an input image and a predicted image, divides the predicted image into a plurality of unit regions and includes a plurality of prediction units included in each unit region. A basic set of predetermined prediction parameters, a classification means for classifying the image into a first group or a second group, and a prediction parameter for designating a prediction image generation method in each prediction unit belonging to the first group. At least a part of the prediction parameters selected by the first selection means, a first selection means for selecting a prediction parameter for specifying a prediction image generation method in each prediction unit belonging to the second group, Select from a reduced set that includes prediction parameters that are less than or equal to the number of prediction parameters included in the basic set. For each prediction unit belonging to the second selection means and the first group, which prediction parameter the first selection means has selected, and for each prediction unit belonging to the second group, And prediction parameter encoding means for encoding which prediction parameter is selected by the second selection means.

  According to the image coding apparatus configured as described above, a prediction parameter that specifies a method of generating a predicted image in each prediction unit belonging to the second group is included in the same unit region as the second group. From a prediction parameter that is a reduced set that includes at least a part of the prediction parameters selected by the first selection unit for each prediction unit belonging to the first group, and is equal to or less than the number of prediction parameters included in the basic set. The reduction set is selected, and the prediction parameter selected by the second selection unit is encoded.

  Here, since the prediction parameter for each prediction unit is generally correlated with the prediction parameter for the prediction unit located in the vicinity of the prediction unit, the prediction parameter selected for each prediction unit belonging to the first group. Is likely to be an appropriate prediction parameter for each prediction unit belonging to the second group. That is, for each prediction unit belonging to the second group, there is a high possibility that the prediction parameter selected from the reduced set is an appropriate prediction parameter. Therefore, according to the above configuration, the prediction parameter can be encoded without reducing the encoding efficiency.

  In the above configuration, the reduced set is a reduced set including at least a part of the prediction parameters selected by the first selection unit, and is equal to or less than the number of prediction parameters included in the basic set. Since the set includes prediction parameters, it is possible to reduce the code amount of information indicating which prediction parameter is selected for each prediction unit belonging to the second group.

  Therefore, according to the above configuration, it is possible to reduce the code amount for designating the prediction parameter without sacrificing the coding efficiency.

  In addition, it is preferable that the reduced set includes only all prediction parameters that are prediction parameters selected by the first selection unit and that are different from each other.

  According to the above configuration, since the number of prediction parameters included in the reduced set can be reduced, the amount of codes can be further reduced. More specifically, since the prediction parameters generally have a spatial correlation, there is a bias in the optimal prediction parameter distribution for each prediction unit in a specific region. Therefore, there is a high probability that only a part of various prediction parameters included in the basic set is used in a specific region. Therefore, the set of prediction parameters selected by the first selection means in a specific region often has fewer elements than the set of prediction parameters included in the basic set. Therefore, the set of all non-overlapping prediction parameters selected by the first selection unit is set as a reduced set, so that the number of elements of the reduced set can be reduced.

  Each of the plurality of unit areas is preferably a coding unit in the image coding apparatus.

  According to the above configuration, the reduced set generation process needs to be executed only once for each coding unit (for example, a macroblock in the H.264 / MPEG-4 AVC standard). The amount can be reduced.

  In the image coding device according to the present invention, the prediction units belonging to the first group and the prediction units belonging to the second group are arranged in a checkered pattern (checkerboard shape). preferable.

  In general, a prediction parameter in each prediction unit has a correlation with a prediction parameter in a neighboring prediction unit.

  According to said structure, since the prediction unit which belongs to the said 1st group and the prediction unit which belongs to the said 2nd group are arrange | positioned in checkered pattern shape (checkerboard shape), it belongs to the said 1st group. The prediction parameter selected for each prediction unit is likely to be an appropriate prediction parameter for each prediction unit belonging to the second group.

  Therefore, according to the above configuration, it is possible to reduce the amount of code necessary for encoding the prediction parameter without sacrificing the encoding efficiency.

  The prediction parameter may specify a prediction mode in intra prediction.

  According to said structure, the code amount required for the encoding of the prediction mode in intra prediction can be reduced, without sacrificing encoding efficiency.

  In addition, the second selection unit is a reduced set including a prediction parameter that defines a method for generating a prediction image on a prediction unit belonging to the second group, including the prediction parameter selected by the first selection unit. It is preferable to select from a reduced set including at least one of a vertical prediction mode, a horizontal prediction mode, and a DC prediction mode in intra prediction.

  In general, the vertical direction prediction mode, the horizontal method prediction mode, and the DC prediction mode are frequently selected prediction modes in intra prediction.

  According to said structure, the said 2nd selection means uses the prediction parameter which selected the prediction parameter which prescribes | regulates the production | generation method of the prediction image on the prediction unit which belongs to the said 2nd group by the said 1st selection means. Can be selected from a reduced set including at least one of a vertical prediction mode, a horizontal prediction mode, and a DC prediction mode in intra prediction, and belongs to the second group. The code amount can be reduced without reducing the prediction accuracy when generating a prediction image on the prediction unit.

  The second selection means selects a prediction parameter from the reduced set when the number of prediction units in a unit region including the first group and the second group is equal to or greater than a predetermined threshold. If not, the prediction parameter may be selected from the basic set.

  According to the above configuration, since the prediction parameter can be selected from the basic set when the number of prediction units in the unit area is less than a predetermined value, the processing amount for selecting the prediction parameter is reduced. can do.

  In the image encoding device according to the present invention, the basic set can be set for each unit region, and the second selection unit is configured to apply to the unit region including the first group and the second group. If the basic set set in this way satisfies a specific condition, a prediction parameter may be selected from the reduced set, and if not, a prediction parameter may be selected from the basic set.

  According to the above configuration, the basic set can be set for each unit area, and the second selection unit is set for a unit area including the first group and the second group. When the basic set satisfies a specific condition, the prediction parameter can be selected from the reduced set, and when not, the prediction parameter can be selected from the basic set. Therefore, the processing amount for selecting the prediction parameter can be reduced. The amount of codes can be reduced while reducing.

  The image coding apparatus according to the present invention is an image coding apparatus that codes a difference between an input image and a predicted image. A prediction parameter that specifies a method for generating a predicted image in each prediction unit is set in the prediction unit. Selection means for selecting from a reduced set including at least a part of a prediction parameter for specifying a prediction image generation method in an encoded prediction unit located in the vicinity, and for each prediction unit, the selection means selects which prediction parameter. It can also be expressed as an image encoding device characterized by including a prediction parameter encoding means for encoding whether or not it is selected.

  In general, a prediction parameter for each prediction unit has a correlation with a prediction parameter for a prediction unit located in the vicinity of the prediction unit. Therefore, the reduced set is most likely to include a prediction parameter in generating a prediction image in the prediction unit. In addition, since the reduced set is composed of at least a part of the prediction parameters for the prediction unit located in the vicinity of the prediction unit, the number of prediction parameters included in the reduced set is for the prediction units other than the prediction unit. The number is smaller than the number of prediction parameters included in the parameter set including the prediction parameters.

  Therefore, the image encoding apparatus according to the present invention can generate encoded data with a small code amount without sacrificing encoding efficiency by adopting the above configuration.

  Further, the prediction parameter encoding means has a code indicating which prediction parameter has been selected by the first selection means as a code indicating which prediction parameter has been selected by the second selection means. It is preferable to use a code with a short code length.

  According to the above configuration, the prediction parameter encoding unit selects which prediction parameter the first selection unit has selected as a code indicating which prediction parameter the second selection unit has selected. Since a code having a shorter code length than the code indicating can be used, information indicating which prediction parameter is selected for each prediction unit belonging to the second group is used with a code with a shorter code length. Can be encoded.

  In addition, the second selection unit is a reduced set including a prediction parameter that defines a method for generating a prediction image on a prediction unit belonging to the second group, including the prediction parameter selected by the first selection unit. It is preferable to select from a reduced set of 2n + 1 (n is an arbitrary natural number) prediction parameters.

  In general, when an element group composed of 2n + 1 elements (n is an arbitrary natural number) is encoded, the compression efficiency is improved as compared with the case where an element group composed of elements other than 2n + 1 elements is encoded.

  According to said structure, the prediction parameter which prescribes | regulates the production | generation method of the prediction image on the prediction unit which belongs to the said 2nd group is a reduction | restoration set containing the prediction parameter selected by the said 1st selection means, Since it is possible to encode which prediction parameter is selected from a reduced set of 2n + 1 (n is an arbitrary natural number) prediction parameters, it is possible to increase the compression efficiency when encoding the prediction parameters. There is a further effect.

  In addition, the image decoding apparatus according to the present invention determines the difference between the original image and the predicted image, which prediction parameter is selected from among a plurality of prediction parameters that specify a predicted image generation method for each prediction unit. In an image decoding apparatus that decodes encoded data obtained by encoding together with selection information that is indicated, a plurality of prediction units included in each of a plurality of unit regions that constitute a prediction image are assigned to a first group or a second group A prediction parameter for specifying a prediction image generation method in each prediction unit belonging to the first group is determined in advance with reference to classification means for classifying into the group and selection information for each prediction unit belonging to the first group. Selection information on each prediction unit belonging to the second group, and a first selection means for selecting from a basic set of prediction parameters obtained A prediction parameter for specifying a prediction image generation method in each prediction unit belonging to the second group is a reduced set including at least a part of the prediction parameters selected by the first selection unit, And a second selection means for selecting from a reduced set of prediction parameters that are equal to or less than the number of prediction parameters included in the basic set.

  According to the image decoding device configured as described above, a prediction parameter that specifies a method for generating a predicted image in each prediction unit belonging to the second group is included in the same unit region as that of the second group. A reduced set including at least a part of the prediction parameters selected by the first selection unit for each prediction unit belonging to one group, and including a prediction parameter equal to or less than the number of prediction parameters included in the basic set. You can choose from a reduced set.

  Here, since the prediction parameter for each prediction unit is generally correlated with the prediction parameter for the prediction unit located in the vicinity of the prediction unit, the prediction parameter selected for each prediction unit belonging to the first group. Is likely to be an appropriate prediction parameter for each prediction unit belonging to the second group. Therefore, according to the above configuration, the prediction parameter can be decoded from the selection information with a smaller code amount without reducing the encoding efficiency.

  In addition, it is preferable that the reduced set includes only all prediction parameters that are prediction parameters selected by the first selection unit and that are different from each other.

  According to the above configuration, since the number of prediction parameters included in the reduced set can be further reduced, the code amount can be further reduced.

  Each of the plurality of unit areas is preferably a decoding unit in the image decoding apparatus.

  According to the above configuration, the generation process of the reduced set needs to be executed only once for each decoding unit (for example, a macroblock in the H.264 / MPEG-4 AVC standard). Can be reduced.

  In the image decoding apparatus according to the present invention, it is preferable that the prediction units belonging to the first group and the prediction units belonging to the second group are arranged in a checkered pattern (checkerboard shape). .

  In general, a prediction parameter in each prediction unit has a correlation with a prediction parameter in a neighboring prediction unit.

  According to the image coding apparatus having a configuration corresponding to the above configuration, the prediction units belonging to the first group and the prediction units belonging to the second group are arranged in a checkered pattern (checkerboard shape). Therefore, an appropriate prediction parameter can be selected for each prediction unit belonging to the second group. Therefore, according to the image encoding apparatus having a configuration corresponding to the above configuration, it is possible to reduce the amount of code necessary for encoding the prediction parameter without sacrificing encoding efficiency.

  According to the image decoding apparatus having the above-described configuration, it is possible to decode the encoded data with the code amount reduced as described above.

  In the image decoding apparatus according to the present invention, the prediction parameter may specify a prediction mode in intra prediction.

  According to said structure, the encoding data by which the code amount of the prediction mode in intra prediction was reduced can be decoded, without sacrificing encoding efficiency.

  In addition, the second selection unit is a reduced set including a prediction parameter that defines a method for generating a prediction image on a prediction unit belonging to the second group, including the prediction parameter selected by the first selection unit. It is preferable to select from a reduced set including at least one of a vertical prediction mode, a horizontal prediction mode, and a DC prediction mode in intra prediction.

  In general, the vertical direction prediction mode, the horizontal method prediction mode, and the DC prediction mode are frequently selected prediction modes in intra prediction.

  According to the image coding apparatus having a configuration corresponding to the above configuration, the second selection unit sets the prediction parameter that defines a method for generating a prediction image on a prediction unit belonging to the second group, to the first parameter. A reduction set including a prediction parameter selected by one selection unit, wherein the reduction set includes at least one of a vertical prediction mode, a horizontal prediction mode, and a DC prediction mode in intra prediction. Therefore, it is possible to reduce the code amount without reducing the prediction accuracy when generating the prediction image on the prediction unit belonging to the second group.

  According to the image decoding apparatus having the above configuration, such encoded data with a small code amount can be decoded.

  The second selection means selects a prediction parameter from the reduced set when the number of prediction units in a unit region including the first group and the second group is equal to or greater than a predetermined threshold. If not, the prediction parameter may be selected from the basic set.

  According to the above configuration, since the prediction parameter can be selected from the basic set when the number of prediction units in the unit area is less than a predetermined value, the processing amount for selecting the prediction parameter is reduced. can do.

  In the image decoding apparatus according to the present invention, the basic set can be set for each unit area, and the second selection unit is arranged in a unit area including the first group and the second group. On the other hand, when the basic set set for the condition satisfies a specific condition, a prediction parameter may be selected from the reduced set, and otherwise, the prediction parameter may be selected from the basic set.

  According to the above configuration, the basic set can be set for each unit area, and the second selection unit is set for a unit area including the first group and the second group. When the basic set satisfies a specific condition, the prediction parameter can be selected from the reduced set, and when not, the prediction parameter can be selected from the basic set. Therefore, the processing amount for selecting the prediction parameter can be reduced. It is possible to decode encoded data with a reduced code amount while reducing the amount of code.

  In addition, the image decoding apparatus according to the present invention determines the difference between the input image and the predicted image, which prediction parameter is selected from among a plurality of prediction parameters that specify a predicted image generation method for each prediction unit. In an image decoding apparatus that decodes encoded data obtained by encoding together with selected selection information, a prediction parameter that specifies a method for generating a predicted image in each prediction unit with reference to the selection information, Expressed as an image decoding apparatus characterized by including selection means for selecting from a reduced set including at least a part of a prediction parameter for specifying a prediction image generation method in a decoded prediction unit located in the vicinity You can also

  In general, a prediction parameter for each prediction unit has a correlation with a prediction parameter for a prediction unit located in the vicinity of the prediction unit. Therefore, the reduced set is most likely to include a prediction parameter in generating a prediction image in the prediction unit. In addition, since the reduced set is composed of at least a part of the prediction parameters for the prediction unit located in the vicinity of the prediction unit, the number of prediction parameters included in the reduced set is for the prediction units other than the prediction unit. The number is smaller than the number of prediction parameters included in the parameter set including the prediction parameters.

  Therefore, an image encoding apparatus having a configuration corresponding to the above configuration can generate encoded data with a small code amount without sacrificing encoding efficiency.

  The image decoding apparatus having the above configuration can decode such encoded data with a small code amount.

  In addition, the second selection unit is a reduced set including a prediction parameter that defines a method for generating a prediction image on a prediction unit belonging to the second group, including the prediction parameter selected by the first selection unit. It is preferable to select from a reduced set of 2n + 1 (n is an arbitrary natural number) prediction parameters.

  In general, when an element group composed of 2n + 1 elements (n is an arbitrary natural number) is encoded, the compression efficiency is improved as compared with the case where an element group composed of elements other than 2n + 1 elements is encoded.

  According to the above configuration, encoded data with high compression efficiency can be decoded in this way.

  Also, the data structure of the encoded data according to the present invention is such that any prediction parameter is selected from among a plurality of prediction parameters that specify the difference between the input image and the prediction image and the prediction image generation method for each prediction unit. A data structure of encoded data obtained by encoding together with selection information indicating whether or not a prediction image is generated by decoding the encoded data, and specifying a prediction image generation method in each prediction unit Including selection information referred to for selecting a parameter from a reduced set including at least a part of a prediction parameter that specifies a method of generating a predicted image in a decoded prediction unit located in the vicinity of the prediction unit; It is characterized by that.

  In general, a prediction parameter for each prediction unit has a correlation with a prediction parameter for a prediction unit located in the vicinity of the prediction unit. Therefore, the reduced set is most likely to include a prediction parameter in generating a prediction image in the prediction unit. In addition, since the reduced set includes at least a part of prediction parameters for a prediction unit located in the vicinity of the prediction unit, the number of prediction parameters specified by the selection information is a prediction unit other than the prediction unit. Is less than the number of prediction parameters included in the parameter set made up of prediction parameters.

  Therefore, the encoded data having the above-described configuration is encoded data with a reduced code amount without sacrificing encoding efficiency.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention can be suitably applied to an image encoding device that encodes an image and generates encoded data, and an image decoding device that decodes encoded data generated using such an image encoding device.

1 Video decoding device 14 MB decoding unit 144 Prediction parameter decoding unit 41 Group determination unit (classification means)
42 switch unit 43 first prediction parameter decoding unit (first selection means)
44 Reduced set derivation unit 45 Second prediction parameter decoding unit (second selection means)
2 Video Encoding Device 24 MB Encoding Unit 242 Prediction Parameter Determination Unit 51 Group Determination Unit (Classification Unit)
52 switch unit 53 first prediction parameter determination unit (first selection means)
54 Reduced set derivation unit 55 Second prediction parameter determination unit (second selection means)
243 prediction parameter encoding unit (prediction parameter encoding means)
248 Predictive image generation unit

Claims (6)

In an image decoding apparatus that decodes encoded data obtained by encoding a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit,
A prediction parameter decoding means for decoding a prediction parameter applied to each prediction unit;
A reduced set construction means for generating a reduced set that is a set of prediction parameters;
Or decodes the first prediction parameters for specifying the prediction parameters included in the basic set consisting of a predetermined prediction parameters, or, the second prediction parameters specifying the prediction parameters included in the reduction set in the reduction set Second prediction parameter decoding means for decoding from a bit string having a length determined by the number of prediction parameters included,
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
Depending on the value of the flag included in the sign-coded data or the first prediction parameters are decoded by the second prediction parameter decoding means or said second prediction parameters according to the value of the flag included in the sign-data -out based on the second prediction parameters are decoded by the decoding means, the prediction parameter decoding means is adapted to decode the prediction parameters, the prediction parameters are a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. The prediction parameter decoding means decodes the prediction parameters included in the reduced set, and the prediction parameter decoding means, when the number of prediction parameters included in the reduced set is 1, the second prediction parameter. by omitting the decoding of the data image decoding apparatus characterized by decoding the prediction parameters. An image decoding method for decoding encoded data obtained by encoding a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit,
A reduced set construction step for generating a reduced set that is a set of prediction parameters applied to each prediction unit ;
A first prediction parameter that identifies a prediction parameter included in a basic set of predetermined prediction parameters is decoded, or a second prediction parameter that identifies a prediction parameter included in the reduction set is included in the reduction set. A second prediction parameter decoding step for decoding from a bit string having a length determined by the number of prediction parameters to be determined;
A prediction parameter decoding step for decoding the prediction parameter ;
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
Depending on the value of the flag included in the sign-coded data or the first prediction parameters are decoded by the second prediction parameter decoding step, or the second prediction parameters according to the value of the flag included in the sign-data based-out in the second prediction parameter decoded in decoding step, in the prediction parameter decoding step, which decodes the prediction parameters, the prediction parameters are a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. In the prediction parameter decoding step, the prediction parameter included in the reduced set is decoded. In the prediction parameter decoding step, when the number of prediction parameters included in the reduced set is 1, the second prediction parameter is decoded. image decoding method characterized by by omitting the decoding of the data decoding the prediction parameters. In an image encoding device that generates encoded data by encoding a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit,
Prediction parameter encoding means for encoding a prediction parameter to be applied to each prediction unit;
A reduced set construction means for generating a reduced set that is a set of prediction parameters;
Or it encodes the first prediction parameters for specifying the prediction parameters included in the basic set consisting of a predetermined prediction parameters, or, the second prediction parameters specifying the prediction parameters included in the reduction set, the reduction set A second prediction parameter encoding means for encoding from a bit string having a length determined by the number of prediction parameters included in
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
Marks Goka are in accordance with the value of the flag included in the encoded data that is encoded by means, or first prediction parameters are coded by the second prediction parameter coding means, or marks Goka is Ru code and depending on the value of the flag included in the data, based-out in the second prediction parameters encoded by the second prediction parameter coding means, said prediction parameter encoding means encodes the prediction parameters The prediction parameter is a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. The prediction parameter encoding means encodes the prediction parameter included in the reduced set, and the prediction parameter encoding means determines that the second parameter when the number of prediction parameters included in the reduced set is one . by omitting the encoding of the prediction parameter image coding apparatus characterized by coding the prediction parameters. In an image encoding method for generating encoded data by encoding a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit,
A reduced set construction step for generating a reduced set that is a set of prediction parameters applied to each prediction unit;
A first prediction parameter that specifies a prediction parameter included in a basic set of predetermined prediction parameters is encoded, or a second prediction parameter that specifies a prediction parameter included in a reduction set is included in the reduction set. A second prediction parameter encoding step for encoding from a bit string having a length determined by the number of prediction parameters included;
A prediction parameter encoding step for encoding the prediction parameter;
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
And depending on the value of the flag included in the encoded Ru encoded data, either the first prediction parameters are encoded in the second prediction parameters coding step, or is included in mark Goka is Ru encoded data and in accordance with the value of the flag,-out based on the second prediction parameters are encoded in the second prediction parameters coding step, in the prediction parameter coding step, which encodes the prediction parameters, the The prediction parameter is a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. In the prediction parameter encoding step, the prediction parameters included in the reduced set are encoded. In the prediction parameter encoding step, when the number of prediction parameters included in the reduced set is 1, the second parameter image encoding method characterized by coding the prediction parameters are omitted encoding of the prediction parameters. Control for causing a computer to execute an image decoding method for decoding encoded data obtained by encoding a difference between an input image and a predicted image together with a prediction parameter used to generate a predicted image for each prediction unit A computer-readable recording medium storing a program,
A reduced set construction step for generating a reduced set that is a set of prediction parameters applied to each prediction unit ;
A first prediction parameter that identifies a prediction parameter included in a basic set of predetermined prediction parameters is decoded, or a second prediction parameter that identifies a prediction parameter included in the reduction set is included in the reduction set. A second prediction parameter decoding step for decoding from a bit string having a length determined by the number of prediction parameters to be determined;
A prediction parameter decoding step for decoding the prediction parameter ;
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
Depending on the value of the flag included in the sign-coded data or the first prediction parameters are decoded by the second prediction parameter decoding step, or the second prediction parameters according to the value of the flag included in the sign-data based-out in the second prediction parameter decoded in decoding step, in the prediction parameter decoding step, which decodes the prediction parameters, the prediction parameters are a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. In the prediction parameter decoding step, the prediction parameter included in the reduced set is decoded. In the prediction parameter decoding step, when the number of prediction parameters included in the reduced set is 1, the second prediction parameter is decoded. decoding the prediction parameters omit decoding of data,
A computer-readable recording medium on which a control program for causing a computer to execute the above is recorded. A control program for causing a computer to execute an image encoding method for generating encoded data by encoding a difference between an input image and a predicted image together with a prediction parameter used for generating a predicted image for each prediction unit A recorded computer-readable recording medium,
A reduced set construction step for generating a reduced set that is a set of prediction parameters applied to each prediction unit;
A first prediction parameter that specifies a prediction parameter included in a basic set of predetermined prediction parameters is encoded, or a second prediction parameter that specifies a prediction parameter included in a reduction set is included in the reduction set. A second prediction parameter encoding step for encoding from a bit string having a length determined by the number of prediction parameters included;
A prediction parameter encoding step for encoding the prediction parameter;
The reduction set with at least one or more prediction parameters included in the neighbor prediction unit area for the target prediction unit, the number of prediction parameters included in the reduction set is less than the number of prediction parameters included in the basic set,
And depending on the value of the flag included in the encoded Ru encoded data, either the first prediction parameters are encoded in the second prediction parameters coding step, or is included in mark Goka is Ru encoded data and in accordance with the value of the flag,-out based on the second prediction parameters are encoded in the second prediction parameters coding step, in the prediction parameter coding step, which encodes the prediction parameters, the The prediction parameter is a motion vector used for motion compensation prediction,
If at least two or more prediction parameters included in the neighborhood prediction unit region for the target prediction unit include a prediction parameter that overlaps between prediction parameters, one of the overlapping prediction parameters is included in the reduced set. In the prediction parameter encoding step, the prediction parameters included in the reduced set are encoded. In the prediction parameter encoding step, when the number of prediction parameters included in the reduced set is 1, the second parameter a computer-readable recording medium a control program to be executed by omitting the encoding of the prediction parameters to encode the prediction parameters into the computer.